HP NonStop SQL/MX

Comparison Guide for

SQL/MP Users

Abstract
This guide describes SQL differences between HP NonStop™ SQL/MP and NonStop
SQL/MX.
Product Version
NonStop SQL/MX Releases 2.0 and 2.1
Supported Release Version Updates (RVUs)
This publication supports G06.23 and all subsequent G-series RVUs until otherwise
indicated by its replacement publication.

Part Number Published

523735-003 August 2005
Document History
Part Number Product Version Published
429827-001 NonStop SQL/MX Release 1.0 March 2001
429857-001 NonStop SQL/MX Release 1.8 December 2002
523735-001 NonStop SQL/MX Release 2.0 June 2004
523735-002 NonStop SQL/MX Release 2.0 August 2004
523735-003 NonStop SQL/MX Release 2.1 August 2005
 HP NonStop SQL/MX
Comparison Guide for SQL/MP
Users

Index Figures Tables

What’s New in This Manual v

Manual Information v
New and Changed Information v
About This Guide vii
Audience vii
Related Documentation vii
Notation Conventions x

1. Comparing the Differences in the Products

Why Build a New DBMS? 1-1
Frequently Asked Questions About Differences 1-2
What’s New in NonStop SQL/MX? 1-4
Redesigned Optimizer and Executor 1-4
Performance 1-4
Parallel Operations for Scalable High Performance 1-4
Manageability 1-5
ANSI Compatibility 1-5
Other Features 1-6

2. Quick Reference to Differences

3. DML Features
Data Types 3-1
Date-time Data Types 3-1
Interval Data Type 3-2
Selecting FRACTION-Only Interval Columns 3-3
Floating-Point Data Types 3-4
File Types 3-4
Identifiers 3-5
Reserved Words 3-5
Literals 3-5
String Delimiters 3-5

Hewlett-Packard Company—523735-003
i
Contents 4. Embedded SQL

Identifiers (continued)
Date-time Literals 3-6
Interval Literals 3-8
Expressions and Functions 3-8
Numeric Value Expressions and Functions 3-9
Datetime Functions 3-9
Interval Value Expressions 3-10
Mathematical Functions 3-10
Sequence Functions 3-10
Aggregate Functions 3-11
Predicates 3-11
LIKE Predicate 3-11
Sort Operations 3-12
SQL DML Statements 3-12
Nonaudited Tables 3-13
INSERT Statement 3-14
SELECT Statement 3-15
UPDATE Statement 3-16
Access Options and Isolation Levels 3-16
Isolation Levels Contrasted Between NonStop SQL/MP and NonStop
SQL/MX 3-16
STABLE Access 3-17
Transaction Management 3-18
Starting and Ending a Transaction 3-18
SET TRANSACTION Statement 3-19
CONTROL TABLE Directive 3-21
CONTROL QUERY Directive 3-23
Query Caching 3-24
Query Type Comparison 3-27
Versioning 3-29

4. Embedded SQL
Host Variable Declarations 4-1
SETSCALE Function 4-1
Variable-Length Character Data in Embedded SQL C Programs 4-2
INVOKE Directive 4-3
Data Type Conversion 4-3
Host Variables With Date-time Data Types 4-3
Indicator Variables 4-3

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

ii
Contents 5. Optimizer and Executor

Embedded SQL Statements 4-4

EXECUTE Statement 4-5
Updatable Cursors 4-6
FETCH Statement 4-6
Positioned UPDATE Statement 4-6
Dynamic SQL 4-6
Descriptor Area 4-7
Error Handling 4-8
Format of Error Codes 4-8
WHENEVER Declarative 4-9
Diagnostics Area 4-9
Statistics Area 4-10
Program Development 4-10
SQL/MP Compiler 4-10
SQL/MX Compiler 4-11
Stored Procedures in Java 4-13
SQL/MP Stored Procedures 4-13
SQL/MX Stored Procedures 4-13
Similarity Checks and Automatic Recompilation 4-13

5. Optimizer and Executor

Optimizer 5-1
Executor 5-3
SQL/MP Iterator Model 5-4
SQL/MX Task Model 5-5
Ways to Influence the Query Execution Plan 5-7
Default Settings 5-7
EXPLAIN Features 5-8
Forcing Query Execution Plans 5-8
Statistics 5-9

6. Data Definition Language (DDL) Differences

SQL Statements 6-1
Database Object Names 6-1
ANSI Names 6-1
Organization of Catalogs and Schemas 6-2
SQL/MX Catalogs 6-2
SQL/MP Catalogs 6-2
Metadata 6-3

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

iii
Contents 7. Utility Differences

Tables 6-4
Clustering Keys 6-4
Table Attributes 6-4
File Types 6-4
Constraints 6-4
Indexes 6-5
Partitions 6-5
Views 6-6
SQL/MX Views 6-6
SQL/MP Views 6-6
Object Security 6-6

7. Utility Differences
DISPLAY USE OF Command 7-1
IMPORT Command 7-1
MXTOOL Utility 7-1
Partition Management: The MODIFY Utility 7-2
POPULATE INDEX Command 7-2
SHOWDDL Command 7-2
SHOWLABEL Command 7-3

Index
Figures
Figure 4-1. SQL/MP Compilation Process 4-11
Figure 4-2. SQL/MX Compilation Process 4-12

Tables
Table 2-1. NonStop SQL/MP Contents Compared to NonStop SQL/MX 2-1
Table 3-1. Support For NonStop SQL/MP DATETIME Literals in NonStop
SQL/MX 3-6
Table 3-2. NonStop SQL/MP and NonStop SQL/MX Functions 3-9
Table 3-3. Mapping NonStop SQL/MP Datetime Functions To NonStop
SQL/MX 3-10
Table 3-4. SQL/MX and SQL/MP DML Statements 3-12
Table 3-5. CONTROL TABLE Directives 3-22
Table 3-6. Query Type Comparison 3-27
Table 4-1. Embedded SQL Statements 4-4

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

iv
 What’s New in This Manual
Manual Information
HP NonStop SQL/MX Comparison Guide for SQL/MP Users

Abstract
This guide describes SQL differences between HP NonStop™ SQL/MP and NonStop
SQL/MX.
Product Version
NonStop SQL/MX Releases 2.0 and 2.1
Supported Release Version Updates (RVUs)
This publication supports G06.23 and all subsequent G-series RVUs until otherwise
indicated by its replacement publication.

Part Number Published

523735-003 August 2005

Document History
Part Number Product Version Published
429827-001 NonStop SQL/MX Release 1.0 March 2001
429857-001 NonStop SQL/MX Release 1.8 December 2002
523735-001 NonStop SQL/MX Release 2.0 June 2004
523735-002 NonStop SQL/MX Release 2.0 August 2004
523735-003 NonStop SQL/MX Release 2.1 August 2005

New and Changed Information

This table summarizes new and changed information for this edition of the manual:
Section Topic New or Changed Information
3 Expressions and Corrected information on CURRENT and EXTEND.
Functions
LIKE Predicate Corrected examples.
4 Similarity Checks and New subsection.
Automatic Recompilation

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

v
What’s New in This Manual New and Changed Information

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

vi
 About This Guide
This guide describes SQL differences between NonStop SQL/MP and NonStop
SQL/MX.
This release of NonStop SQL/MX enables applications to use the SQL/MX high-
performance database compiler and executor to access SQL/MP and SQL/MX
databases. You access tables, views, indexes, and catalogs by using SQL/MX Data
Manipulation Language (DML) statements.

Audience
This guide is intended for database administrators and application programmers who
want to:
• Understand how NonStop SQL/MX differs from NonStop SQL/MP.
• Write new SQL/MX applications.
• Write new SQL/MP applications that will be ported to NonStop SQL/MX in the
future.
• Investigate migrating existing SQL/MP applications to NonStop SQL/MX.
Learning about the features of SQL implemented in NonStop SQL/MX enables you, as
a database administrator or an application programmer, to choose the appropriate
database for your needs. In particular, it is important for those who want to rework
SQL/MP applications to NonStop SQL/MX to begin thinking about specific differences.

Related Documentation
This manual is part of the NonStop SQL/MX library of manuals, which includes:
Introductory Guides
SQL/MX Comparison Guide Describes SQL differences between NonStop
for SQL/MP Users SQL/MP and NonStop SQL/MX.
SQL/MX Quick Start Describes basic techniques for using SQL in the
SQL/MX conversational interface (MXCI). Includes
information about installing the sample database.
Reference Manuals
SQL/MX Reference Manual Describes the syntax of SQL/MX statements, MXCI
commands, functions, and other SQL/MX language
elements.
SQL/MX Connectivity Describes the SQL/MX administrative command
Service Administrative library (MACL) available with the SQL/MX
Command Reference conversational interface (MXCI).
DataLoader/MX Reference Describes the features and functions of the
Manual DataLoader/MX product, a tool to load SQL/MX
databases.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

vii
About This Guide Related Documentation

SQL/MX Messages Manual Describes SQL/MX messages.

SQL/MX Glossary Defines SQL/MX terminology.
Programming Manuals
SQL/MX Programming Describes how to embed SQL/MX statements in
Manual for C and COBOL ANSI C and COBOL programs.
SQL/MX Programming Describes how to embed SQL/MX statements in
Manual for Java Java programs according to the SQLJ standard.
Specialized Guides
SQL/MX Installation and Describes how to plan, install, create, and manage
Management Guide an SQL/MX database. Explains how to use
installation and management commands and
utilities.
SQL/MX Query Guide Describes how to understand query execution
plans and write optimal queries for an SQL/MX
database.
SQL/MX Data Mining Guide Describes the SQL/MX data structures and
operations to carry out the knowledge-discovery
process.
SQL/MX Queuing and Describes how SQL/MX integrates transactional
Publish/Subscribe Services queuing and publish/subscribe services into its
database infrastructure.
SQL/MX Report Writer Guide Describes how to produce formatted reports using
data from a NonStop SQL/MX database.
SQL/MX Connectivity Describes how to install and manage the SQL/MX
Service Manual Connectivity Service (MXCS), which enables
applications developed for the Microsoft Open
Database Connectivity (ODBC) application
programming interface (API) and other connectivity
APIs to use SQL/MX.
SQL/MX Guide to Stored Describes how to use stored procedures that are
Procedures in Java written in Java within SQL/MX.
Online Help
The SQL/MX Online Help consists of:
Reference Help Overview and reference entries from the SQL/MX
Reference Manual.
Messages Help Individual messages grouped by source from the
SQL/MX Messages Manual.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

viii
About This Guide Related Documentation

Glossary Help Terms and definitions from the SQL/MX Glossary.

NSM/web Help Context-sensitive help topics that describe how to
use the NSM/web management tool.
Visual Query Planner Help Context-sensitive help topics that describe how to
use the Visual Query Planner graphical user
interface.
The NSM/web and Visual Query Planner help systems are accessible from their
respective applications. You can download the Reference, Messages, and Glossary
online help from the $SYSTEM.ZMXHELP subvolume or from the HP NonStop
Technical Library (NTL). For more information about downloading online help, see the
SQL/MX Installation and Management Guide.
These manuals are part of the SQL/MP library of manuals and are essential references
for information about SQL/MP Data Definition Language (DDL) and SQL/MP
installation and management:
Related SQL/MP Manuals
SQL/MP Reference Manual Describes the SQL/MP language elements,
expressions, predicates, functions, and statements.
SQL/MP Installation and Describes how to plan, install, create, and manage
Management Guide an SQL/MP database. Describes installation and
management commands and SQL/MP catalogs
and files.

This documentation is helpful for understanding the concepts and terminology of this
manual:
Other Related Documentation
C/C++ Programmer’s Guide Describes HP extensions to the C and C++
languages for NonStop servers.
SQL/MX Connectivity Describes how to install and manage MXCS, which
Service Manual enables applications developed for Microsoft Open
Database Connectivity (ODBC) application
programming interface (API) to use SQL/MX.
Extensible Markup Language Describes the W3 standards for XML at
(XML) Web site http://www.w3.org/XML.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

ix
About This Guide Notation Conventions

This figure shows the manuals in the SQL/MX library:

Introductory Guides Programming Manuals

SQL/MX SQL/MX SQL/MX SQL/MX SQL/MX

Comparison Quick Start Programming Programming Guide to
Guide for Manual for C Manual for Stored
SQL/MP and COBOL Java Procedures
Users in Java

Reference Manuals

SQL/MX SQL/MX SQL/MX SQL/MX Data-

Reference Messages Glossary Connectivity Loader/MX
Manual Manual Service Reference
Command Manual
Reference

Specialized Guides SQL/MX Online Help

Reference Messages
SQL/MX SQL/MX SQL/MX Help Help
Installation Query Guide Data Mining
and Guide
Management
Guide

SQL/MX SQL/MX Glossary NSM/web

SQL/MX Help
Queuing Report Connectivity Help
and Publish/ Writer Service
Subscribe Guide Manual
Services

VST001.vsd

Notation Conventions
Hypertext Links
Blue underline is used to indicate a hypertext link within text. By clicking a passage of
text with a blue underline, you are taken to the location described. For example:
This requirement is described under Backup DAM Volumes and Physical Disk
Drives on page 3-2.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

x
About This Guide General Syntax Notation

General Syntax Notation

This list summarizes the notation conventions for syntax presentation in this manual.

UPPERCASE LETTERS. Uppercase letters indicate keywords and reserved words. Type
these items exactly as shown. Items not enclosed in brackets are required. For
example:
MAXATTACH

lowercase italic letters. Lowercase italic letters indicate variable items that you supply.
Items not enclosed in brackets are required. For example:
file-name

computer type. Computer type letters within text indicate C and Open System Services
(OSS) keywords and reserved words. Type these items exactly as shown. Items not
enclosed in brackets are required. For example:
myfile.c

italic computer type. Italic computer type letters within text indicate C and Open
System Services (OSS) variable items that you supply. Items not enclosed in brackets
are required. For example:
pathname

[ ] Brackets. Brackets enclose optional syntax items. For example:

TERM [\system-name.]$terminal-name
INT[ERRUPTS]
A group of items enclosed in brackets is a list from which you can choose one item or
none. The items in the list can be arranged either vertically, with aligned brackets on
each side of the list, or horizontally, enclosed in a pair of brackets and separated by
vertical lines. For example:
FC [ num ]
[ -num ]
[ text ]
K [ X | D ] address

{ } Braces. A group of items enclosed in braces is a list from which you are required to
choose one item. The items in the list can be arranged either vertically, with aligned
braces on each side of the list, or horizontally, enclosed in a pair of braces and
separated by vertical lines. For example:
LISTOPENS PROCESS { $appl-mgr-name }
{ $process-name }
ALLOWSU { ON | OFF }

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

xi
About This Guide General Syntax Notation

| Vertical Line. A vertical line separates alternatives in a horizontal list that is enclosed in
brackets or braces. For example:
INSPECT { OFF | ON | SAVEABEND }

… Ellipsis. An ellipsis immediately following a pair of brackets or braces indicates that you
can repeat the enclosed sequence of syntax items any number of times. For example:
M address [ , new-value ]…
[ - ] {0|1|2|3|4|5|6|7|8|9}…
An ellipsis immediately following a single syntax item indicates that you can repeat that
syntax item any number of times. For example:
"s-char…"

Punctuation. Parentheses, commas, semicolons, and other symbols not previously

described must be typed as shown. For example:
error := NEXTFILENAME ( file-name ) ;
LISTOPENS SU $process-name.#su-name
Quotation marks around a symbol such as a bracket or brace indicate the symbol is a
required character that you must type as shown. For example:
"[" repetition-constant-list "]"

Item Spacing. Spaces shown between items are required unless one of the items is a
punctuation symbol such as a parenthesis or a comma. For example:
CALL STEPMOM ( process-id ) ;
If there is no space between two items, spaces are not permitted. In this example, no
spaces are permitted between the period and any other items:
$process-name.#su-name

Line Spacing. If the syntax of a command is too long to fit on a single line, each
continuation line is indented three spaces and is separated from the preceding line by
a blank line. This spacing distinguishes items in a continuation line from items in a
vertical list of selections. For example:
ALTER [ / OUT file-spec / ] LINE

[ , attribute-spec ]…

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

xii
 1
Comparing the Differences in the
Products
NonStop SQL/MX is HP’s next-generation relational database management system
designed for business-critical applications. SQL/MX software brings traditional
NonStop fundamentals—high availability, scalability, reliability, and parallel
processing—to a distributed database.
NonStop SQL/MP is HP’s relational database management system that was first
introduced in 1987. Since then, many new features and performance enhancements
have been added. NonStop SQL/MP is widely used today for critical OLTP and DSS
applications running on HP NonStop servers.

Why Build a New DBMS?

NonStop SQL/MX builds upon the heritage and the many years of experience gained
with NonStop SQL/MP. Building on the capabilities of NonStop SQL/MP, ANSI SQL
standard (1992) features have been added that enable you to create applications for
open systems. Parallel query capabilities and ease of management have made
NonStop SQL/MP the preferred database for large business intelligence and
Operational Data Store (ODS) applications. In addition, NonStop SQL/MP has been the
solution for business-critical OLTP applications, such as securities trading and stock
exchange activities.
NonStop SQL/MP has become the choice for large OLTP, business intelligence, and
operational data store applications over the past 12 years. Businesses such as
securities trading, stock exchanges, retail, and manufacturing rely on NonStop
SQL/MP to manage their daily business, because they require highly available,
reliable, and scalable features. Over the years, NonStop SQL/MP has been enhanced
with features and is well tuned for both OLTP and DSS environments. HP recognizes
the importance of NonStop SQL/MP and will continue to support it.
NonStop SQL/MX uses SQL/MP’s capabilities and extends them to include data mining
abilities. NonStop SQL/MX inherits those same performance and high-availability
characteristics and provides a familiar environment for SQL/MP database
professionals by building on the features NonStop SQL/MP currently offers.
So, why this new database management system? NonStop SQL/MP was conceived in
1987, long before the demands of the internet and the needs of e-business. NonStop
SQL/MX was created to support the newer ANSI standard, which has evolved
considerably since NonStop SQL/MP was first released. Customer needs, relating to
scalability, portability, memory, and processor demands, have all contributed to the
need to provide a new database. NonStop SQL/MX meets the needs of both the
internet and the e-business market.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

1 -1
Comparing the Differences in the Products Frequently Asked Questions About Differences

Frequently Asked Questions About

Differences
The following list of questions about the interaction between NonStop SQL/MP and
NonStop SQL/MX was generated from customer response at a recent user
conference.
• Will customers be forced to migrate applications from NonStop SQL/MP to
NonStop SQL/MX?
No. The data-sharing architecture allows you to build new applications that take
advantage of the new SQL/MX features while accessing data maintained in
existing SQL/MP applications. You can choose which pieces of existing
applications will be coded for NonStop SQL/MX and which pieces will continue to
use NonStop SQL/MP. NonStop SQL/MX and NonStop SQL/MP can run
concurrently against the same data for both read and write. A few SQL/MP data
types are not supported in NonStop SQL/MX. See Section 3, DML Features for a
list of data types supported by NonStop SQL/MX.
• Will the new features of NonStop SQL/MX, such as GRANT/REVOKE, be made
available in NonStop SQL/MP?
No. The SQL/MP architecture does not comply with the ANSI-92 standard.
• What are future plans for NonStop SQL/MP? Will SQL/MP be enhanced with new
features?
NonStop SQL/MP was designed in the mid-1980’s using the available technology.
The SQL standard has evolved since then. Although support will continue for
NonStop SQL/MP, most new features will be implemented only in NonStop
SQL/MX because the architecture of NonStop SQL/MX facilitates adding those
new features.
• Do you need to change your SQL/MP tables to run NonStop SQL/MX?
In many cases, yes. Some SQL/MP features do not map directly to NonStop
SQL/MX. For descriptions of SQL/MP features that do not map to NonStop
SQL/MX, see specific entries in the SQL/MX Reference Manual.
• Is dynamic SQL easier to code in NonStop SQL/MX than in NonStop SQL/MP?
Yes. NonStop SQL/MX uses the ANSI model for dynamic SQL. You use SET
DESCRIPTOR and GET DESCRIPTOR instead of SQLCA. You use GET
DIAGNOSTICS for error handling instead of SQLDA. See Section 4, Embedded
SQL, for additional information about dynamic SQL. In addition, see the SQL/MX
Reference Manual and the SQL/MX Programming Guide for C and COBOL.
• Can you use DEFINEs in NonStop SQL/MX?

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

1 -2
Comparing the Differences in the Products Frequently Asked Questions About Differences

Yes, but DEFINEs for table naming are not necessary in NonStop SQL/MX.
DEFINES can be used only for SQL/MP objects referenced through the SQL/MX
application. SQL/MX objects cannot be referenced with DEFINES.
In addition, DEFINEs must be available to NonStop SQL/MX at run time, or you will
receive an error. NonStop SQL/MP does not have this requirement.
• Does NonStop SQL/MX support similarity checking and execution-time name
resolution?
Yes. NonStop SQL/MX supports similarity checking and execution-time name
resolution with some limitations.
• Does NonStop SQL/MX support Report Writer?
Yes. In addition, you can use third-party tools for reports. If you need assistance in
choosing a tool, contact your HP representative.
• Can NonStop SQL/MX run under both the HP NonStop Kernel Open System
Services (OSS) and Guardian Services environments?
Many key SQL/MX processes run as OSS processes (as do HP NonStop Tuxedo,
HP NonStop DOM, and Java processes). Your applications can be run on
Guardian with an OSS pass-through command. See the SQL/MX Programming
Guide for C and COBOL for more information.
• What benefits will NonStop SQL/MX provide for OLTP applications?
You can now build your OLTP application by using ANSI-standard SQL. This
feature allows you to leverage the expertise that your developers have gained in
working on other ANSI-standard databases and quickly transfer those skills to
SQL/MX development. It also allows you to build applications that can be ported
between DBMS systems with a minimum of change. Several new SQL/MX
features, including rowsets and compound statements, provide significant
performance improvements for OLTP applications. NonStop SQL/MX also supports
a publish and subscribe feature that supports transactional-based queuing.
• Does TACL support NonStop SQL/MX?
Yes. TACL supports SQL/MX objects when the Guardian name of the object is
specified. TACL commands do not accept ANSI names as input and do not return
ANSI names in their output. These TACL commands support SQL/MX objects:
° FILEINFO
° FILENAMES
° FILES
° #FILEGETLOCKINFO
° #FILEINFO
° #FILENAMES
° #LOCKINFO
° #OPENINFO
° #XFILEINFO

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

1 -3
Comparing the Differences in the Products What’s New in NonStop SQL/MX?

° #XFILENAMES
° #XFILES
In general, the commands work the same and return the same information for
SQL/MX objects as they do for non-SQL/MX objects. Exceptions are noted in the
TACL Reference Manual.
These commands return errors when used with SQL/MX objects and are not
supported:

° COPYDUMP
° PURGE
° RENAME
° #PURGE
° #RENAME

What’s New in NonStop SQL/MX?

Redesigned Optimizer and Executor
NonStop SQL/MX uses a redesigned optimizer and executor. The top-down, cost-
based, rules-driven optimizer considers more plans than NonStop SQL/MP.
Extensibility is provided through the compiler, because new transformations using rules
can be added with ease to the optimizer. The redesigned executor also provides
extensibility that uses a data-flow architecture. The executor takes advantage internally
of completely no-waited operations and provides for more work in the HP NonStop
Data Access Manager (DAM), including multiple forms of parallelism and scalability on
multiple levels.
For additional information about differences between the SQL/MP and SQL/MX
optimizer features, see Section 5, Optimizer and Executor.

Performance
In addition to considering many more plans, the SQL/MX optimizer allows you to push
down certain plans that involve compound statements or nested joins into the Data
Access Manager. You can influence a query plan by rewriting portions of a plan with
the CONTROL QUERY SHAPE command. NonStop SQL/MX provides histogram
statistics that provide more accurate statistics of the data in the database. These
statistics are also used by the optimizer to create efficient query plans.
For additional information about optimizer features, see Section 5, Optimizer and
Executor.

Parallel Operations for Scalable High Performance

NonStop SQL/MX can execute large queries that scan large portions of the database in
parallel. Although NonStop SQL/MX supports three basic forms of parallelism,
partitioned parallelism is the main type of parallelism used. Partitioned parallelism is

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

1 -4
Comparing the Differences in the Products Manageability

the ability to divide the data to be processed into partitions and work on each partition
in parallel. In a partitioned parallel plan, multiple operators all work on the same plans,
and results are merged by using multiple pipelines. Although NonStop SQL/MP uses
some partitioned parallelism to process in parallel, NonStop SQL/MX has been
designed to take advantage of both executor server process (ESP) parallelism and
Data Access Manager parallelism.
For additional information about parallelism, see the SQL/MX Query Guide.

Manageability
For parallelism in NonStop SQL/MP, one ESP per partition is started. NonStop SQL/MX
simplifies ESP management by starting only the number of ESPs it needs to process a
plan in parallel if the optimizer determines that ESP partitioned parallelism will provide
the best priced query plan.
NonStop SQL/MX provides a number of user-controlled defaults that allow you to
change default settings that affect optimization, performance, and the generation of
histogram statistics.
Like NonStop SQL/MP, NonStop SQL/MX provides EXPLAIN output. However,
NonStop SQL/MX provides EXPLAIN as a table, and the output is in machine-readable
format. You can easily extract the EXPLAIN output from embedded SQL programs. In
addition, NonStop SQL/MX provides a GUI-based query tool called Visual Query
Planner that graphically displays the results of query plans.

ANSI Compatibility
NonStop SQL/MX is open, portable, and, most important, uses an industry standard:
the American National Standards Institute (ANSI) version of SQL that was created in
1992. This version of SQL is referred to as SQL-92—a standard that is considered far
more complete than earlier versions of SQL for business-critical applications. SQL-92
was replaced by SQL:1999. Core SQL:1999 contains all of Entry SQL-92, plus much of
Intermediate SQL-92 and some of Full SQL-92, plus a few new SQL:1999 features.
NonStop SQL/MX is based on SQL:1999 and includes support for features from higher
levels of the standard. NonStop SQL/MP is based on the SQL-89 standard.
Applications with SQL statements that adhere to the SQL:1999 standard can be ported
to run against a variety of database management systems, including NonStop
SQL/MX. If you are trained to write applications with other SQL-92 compliant database
management systems, you can easily apply your skills to applications for NonStop
SQL/MX.
For details about how NonStop SQL/MX features comply with ANSI standards, see the
SQL/MX Reference Manual.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

1 -5
Comparing the Differences in the Products Other Features

Other Features
Many other features have been added to NonStop SQL/MX that are not discussed in
this guide. Some of those features include:
• Data mining
• Publish and subscribe
• Compound statements and rowsets
For additional information regarding these topics, see the SQL/MX Data Mining Guide,
the SQL/MX Queuing and Publish/Subscribe Services, and the SQL/MX Programming
Guide for C and COBOL.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

1 -6
2 Quick Reference to Differences
Table 2-1 lists the content of NonStop SQL/MP as found in the SQL/MP Reference
Manual, the SQL/MX Reference Manual, and other manuals. The column
Implemented In NonStop SQL/MX? indicates whether the item is supported in
NonStop SQL/MX (yes, no, or changed.) “Changed” means that the item is supported
differently in NonStop SQL/MX than it was in NonStop SQL/MP.
NonStop SQL/MX supports Report Writer. In the table, Report Writer commands are
identified in the Comments column.
Remember that NonStop SQL/MX has many features that are not listed here because
this list is NonStop SQL/MP-focused. See the NonStop SQL/MX manuals, especially
the SQL/MX Reference Manual, for further information.

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 1 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
Access option Changed See Section 3, DML Features
ADD DEFINE Yes See SQL/MX Reference Manual
Aggregate function Yes See Section 3, DML Features and
the SQL/MX Reference Manual
Alias Changed Used differently in each product.
See SQL/MP Reference Manual and
SQL/MX Reference Manual
ALLOCATE file attribute Yes See SQL/MX Reference Manual
ALTER CATALOG statement No
ALTER COLLATION No
statement
ALTER DEFINE statement Yes See SQL/MX Reference Manual
ALTER INDEX statement Yes Fewer capabilities. See SQL/MX
Reference Manual
ALTER PROGRAM No
statement
ALTER TABLE statement Yes Fewer capabilities. See SQL/MX
Reference Manual
ALTER VIEW statement No
APPEND command No
APPENDCANCEL command No

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2 -1
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 2 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
APPENDRESTART No
command
ASCII character set Yes
AS clause Yes See SQL/MX Report Writer Guide
AS DATE/TIME clause Yes See SQL/MX Report Writer Guide
AUDIT file attribute No See SQL/MP Reference Manual
AUDITCOMPRESS file Yes See SQL/MX Reference Manual
attribute
Audited tables Yes See SQL/MX Reference Manual
AVG function Yes See SQL/MX Reference Manual
BACKUP utility Yes See SQL/MP Reference Manual
BASETABS table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
BEGIN DECLARE SECTION Yes See Section 4, Embedded SQL and
statement SQL/MX Reference Manual
BEGIN WORK statement Yes See Section 3, DML Features and
SQL/MX Reference Manual
BETWEEN predicate Yes See SQL/MX Reference Manual
BLOCKSIZE file attribute Yes Fewer capabilities. See SQL/MX
Reference Manual
BREAK FOOTING command Yes See SQL/MX Report Writer Guide
BREAK ON command Yes See SQL/MX Report Writer Guide
BREAK TITLE command Yes See SQL/MX Report Writer Guide
BUFFERED file attribute No See SQL/MP Reference Manual
CANCEL command Yes See SQL/MX Report Writer Guide
CASE expression Yes See SQL/MX Reference Manual
CAST function Yes See Section 3, DML Features and
SQL/MX Reference Manual
CATALOG command No See SQL/MP Reference Manual
Catalogs Changed See Organization of Catalogs and
Schemas on page 6-2

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2 -2
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 3 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
CATALOGS table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
CENTER_REPORT option Yes See SQL/MX Report Writer Guide
Character data type Yes See SQL/MX Reference Manual
Character expression Yes See SQL/MX Reference Manual
Character sets Yes See SQL/MX Reference Manual
CHAR_LENGTH function Yes See SQL/MX Reference Manual
CLEANUP command No
CLEARONPURGE file Yes See SQL/MX Reference Manual
attribute
CLOSE statement Yes See SQL/MX Reference Manual
Clustering keys Yes See SQL/MX Reference Manual
COLLATE clause Changed Can only be specified as COLLATE
DEFAULT. See SQL/MX Reference
Manual
Collation definitions No
Collations Yes Fewer capabilities. See SQL/MX
Reference Manual
Column Identifier No See SQL/MP Reference Manual
Columns Yes See SQL/MX Reference Manual
COLUMNS table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
COMMENT statement No
Comments Yes See SQL/MX Reference Manual
COMMENTS table No See SQL/MP Reference Manual
COMMIT option No
COMMIT WORK statement Yes See SQL/MX Reference Manual
Comparison predicate Yes See SQL/MX Reference Manual
COMPUTE_TIMESTAMP No
function
CONCAT clause Yes See SQL/MX Report Writer Guide

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2 -3
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 4 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
Concurrency Yes See SQL/MX Reference Manual
Constraints Yes In CREATE TABLE and ALTER
TABLE statements. See SQL/MX
Reference Manual
CONSTRNT table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
CONTINUE statement No
CONTROL EXECUTOR Changed Parallelism enabled through
directive SYSTEM_DEFAULTS table. See
Section 5, Optimizer and Executor
CONTROL QUERY directive Changed Implemented as CONTROL QUERY
DEFAULT. See Section 3, DML
Features and SQL/MX Reference
Manual
CONTROL TABLE directive Changed See Section 3, DML Features and
SQL/MX Reference Manual
CONVERT command No See SQL/MP Reference Manual
CONVERTTIMESTAMP Yes See SQL/MX Reference Manual
function
COPY command No See SQL/MP Reference Manual
Correlation names Yes See SQL/MX Reference Manual
COUNT function Yes See SQL/MX Reference Manual
CPRLSRCE table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
CPRULES table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
CREATE CATALOG Changed No security capabilities. See
statement SQL/MX Reference Manual
CREATE COLLATION No See SQL/MP Reference Manual
statement

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2 -4
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 5 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
CREATE CONSTRAINT Changed Performed with CREATE TABLE
statement statement. See SQL/MX Reference
Manual
CREATE INDEX statement Yes Some differences. See SQL/MX
Reference Manual
CREATE SYSTEM Changed See Organization of Catalogs and
CATALOG statement Schemas on page 6-2 and SQL/MX
Reference Manual
CREATE TABLE statement Yes Some differences. See SQL/MX
Reference Manual
CREATE VIEW statement Changed Some differences. See Views on
page 6-6 and SQL/MX Reference
Manual
CURRENT function Yes See SQL/MX Reference Manual
CURRENT_TIMESTAMP Yes See SQL/MX Reference Manual
function
Cursors Yes See SQL/MX Programming Guide
for C and COBOL and SQL/MX
Reference Manual
c89 Yes See SQL/MX Programming Guide
for C and COBOL
Data Dictionary Yes Metadata in NonStop SQL/MX. See
SQL/MX Reference Manual
Data types Yes See SQL/MX Reference Manual
DATE data type Yes See SQL/MX Reference Manual
DATE_FORMAT option Yes See SQL/MX Report Writer Guide
Date-time data types Changed See Section 3, DML Features
Date-time functions Changed See Section 3, DML Features
Date-time literals Changed See Section 3, DML Features
DATEFORMAT function Yes See SQL/MX Reference Manual
DATETIME data type Changed See Section 3, DML Features and
SQL/MX Reference Manual
DAYOFWEEK function Yes See SQL/MX Reference Manual

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2 -5
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 6 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
DCL (Data Control Changed See SQL/MX Reference Manual
Language) statements
Deadlocks Yes See Section 3, DML Features and
SQL/MX Reference Manual
(CONTROL TABLE)
DEALLOCATE file attribute Yes See SQL/MX Reference Manual
DECIMAL_POINT option Yes See SQL/MX Report Writer Guide
DECLARE CURSOR Yes See SQL/MX Programming Guide
statement for C and COBOL and SQL/MX
Reference Manual
DCOMPRESS No See SQL/MP Reference Manual
DEFAULT clause Yes See SQL/MX Reference Manual
DEFINEs Yes See SQL/MX Reference Manual
DELETE DEFINE statement Yes See SQL/MX Reference Manual
DELETE statement Yes Some differences. See SQL/MX
Reference Manual and see SQL/MX
Programming Guide for C and
COBOL
DESCRIBE INPUT Yes See SQL/MX Programming Guide
statement for C and COBOL and SQL/MX
Reference Manual
DESCRIBE statement Yes See SQL/MX Reference Manual
Detail alias No
DETAIL command Yes See SQL/MX Report Writer Guide
DISPLAY STATISTICS Changed See SQL/MX Reference Manual
command
DISPLAY USE OF command Changed Differently implemented. See
DISPLAY USE OF Command on
page 7-1, Some functions replaced
by mxtool INFO command
DISTINCT clause Yes See SQL/MX Reference Manual
Data Manipulation Language Yes See Section 3, DML Features and
(DML) statements SQL/MX Reference Manual
DOWNGRADE CATALOG No See SQL/MP Reference Manual
command

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2 -6
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 7 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
DOWNGRADE SYSTEM No See SQL/MP Reference Manual
CATALOG command
DROP statement Changed Replaced by specific DROP object
statements. See SQL/MX Reference
Manual
DROP SYSTEM CATALOG Changed Replaced by DROP SQL statement.
statement See SQL/MX Reference Manual
DSL (Data Status Language) No
statements
DSLACK file attribute No See SQL/MP Reference Manual
DUP command Yes See SQL/MX Reference Manual
Dynamic SQL Yes See Section 4, Embedded SQL and
SQL/MX Programming Guide for C
and COBOL
EDIT command No
Embedded SQL Yes See Section 4, Embedded SQL,
SQL/MX Reference Manual,
SQL/MX Programming Guide for C
and COBOL, and SQL/MX
Programming Guide for C and
COBOL.
END DECLARE SECTION Yes See SQL/MX Reference Manual
statement
ENV command Yes See SQL/MX Reference Manual
ERROR command Yes See SQL/MX Reference Manual
Error messages Yes See SQL/MX Reference Manual
EXECUTE statement Yes See SQL/MX Reference Manual
EXECUTE IMMEDIATE Yes See SQL/MX Reference Manual
statement
EXISTS predicate Yes See SQL/MX Reference Manual
EXIT command Yes See SQL/MX Reference Manual
EXPLAIN directive Changed See Section 5, Optimizer and
Executor, SQL/MX Query Guide,
and SQL/MX Reference Manual
Expressions Yes See SQL/MX Reference Manual

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2 -7
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 8 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
EXTEND function No
EXTENT file attribute Yes See SQL/MX Reference Manual
FC command Yes See SQL/MX Reference Manual
FETCH statement Yes See SQL/MX Reference Manual
File attributes Changed See SQL/MP Reference Manual
File organization Changed See SQL/MP Reference Manual
FILEINFO command Changed Replaced SHOWLABEL utility. See
SQL/MX Reference Manual.
FILENAMES command No
FILES command No
FILES table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
Filesets No
Floating point data type for C Yes See the SQL/MX Reference Manual
and C++
FREE RESOURCES No
statement
Functions Yes See Section 3, DML Features and
SQL/MX Reference Manual
FUP command No
Generalized owner No
GET CATALOG OF No
SYSTEM statement
GET VERSION OF No
PROGRAM statement
GET VERSION statement No
GOAWAY command Changed Replaced by mxtool FIXUP
command
Group manager No
Guardian names Yes See the SQL/MX Reference Manual
HEADING clause Yes See SQL/MX Reference Manual
HEADINGS option Yes See SQL/MX Report Writer Guide

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2 -8
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 9 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
HELP command Changed HTML Help will be provided
HELP TEXT statement No
HISTORY command Yes See SQL/MX Reference Manual
Host identifiers Yes See SQL/MX Programming Guide
for C and COBOL
Host programs Yes See SQL/MX Programming Guide
for C and COBOL
Host variables Yes See Section 4, Embedded SQL and
SQL/MX Programming Guide for C
and COBOL
ICOMPRESS file attribute No
IF/THEN/ELSE clause Yes See SQL/MX Report Writer Guide
IN predicate Yes See SQL/MX Reference Manual
INCLUDE SQLCA directive Changed Use diagnostics area for error
handling. See Section 4, Embedded
SQL and SQL/MX Programming
Guide for C and COBOL
INCLUDE SQLDA directive Changed Use descriptor area for error
handling. See Section 4, Embedded
SQL and SQL/MX Programming
Guide for C and COBOL
INCLUDE SQLSA directive No
INCLUDE STRUCTURES No
directive
Index keys Yes See SQL/MX Reference Manual
INDEXES table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
Indicator variables and Yes See Section 4, Embedded SQL and
Indicator parameters SQL/MX Programming Guide for C
and COBOL
INFO DEFINE statement Yes See SQL/MX Reference Manual
INITIALIZE SQL command Yes See SQL/MX Reference Manual

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2 -9
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 10 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
INSERT statement Yes See SQL/MX Reference Manual and
SQL/MX Programming Guide for C
and COBOL
INTERVAL data type Yes See SQL/MX Reference Manual
INTERVAL literal Yes See SQL/MX Reference Manual
INVOKE directive and Changed Some functions replaced by
command SQL/MX SHOWDDL command. See
Section 4, Embedded SQL, SQL/MX
Programming Guide for C and
COBOL and SQL/MX Reference
Manual
ISLACK file attribute No
Joins Yes See SQL/MX Reference Manual
JULIANTIMESTAMP function Yes See SQL/MX Reference Manual
Keys Yes See SQL/MX Reference Manual
KEYS table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
LEFT_MARGIN option Yes See SQL/MX Report Writer Guide
LIKE predicate Yes See SQL/MX Reference Manual
Limits Yes See SQL/MX Reference Manual
LINE_NUMBER function Yes See SQL/MX Report Writer Guide
LINE_SPACING option Yes See SQL/MX Report Writer Guide
LIST command Yes See SQL/MX Report Writer Guide
Literals Yes See Section 3, DML Features and
SQL/MX Reference Manual
LOAD command Changed Replaced by IMPORT utility. See
SQL/MX Reference Manual
LOCK TABLE statement Yes See SQL/MX Reference Manual
Locking Yes See SQL/MX Reference Manual
LOCKLENGTH file attribute No
LOG command Yes See SQL/MX Reference Manual
LOGICAL_FOLDING option Yes See SQL/MX Report Writer Guide

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2- 10
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 11 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
MAX function Yes See SQL/MX Reference Manual
MAXEXTENTS file attribute Yes See SQL/MX Reference Manual
Message file Changed See SQL/MX Online Help (HTML
format) for messages and online
help text
MIN function Yes See SQL/MX Reference Manual
MODIFY CATALOG No
command
MODIFY LABEL command No
MODIFY REGISTER No
command
Multibyte character sets Yes
NAME command Yes See SQL/MX Report Writer Guide
NAME option No
Name resolution Yes See SQL/MX Reference Manual and
SQL/MX Programming Guide for C
and COBOL
Names Yes See SQL/MX Reference Manual
NEWLINE_CHAR option Yes See SQL/MX Report Writer Guide
Nonaudited tables No
NOPURGEUNTIL file No
attribute
NULL predicate Yes See SQL/MX Reference Manual
Null values Yes See SQL/MX Reference Manual
NULL_DISPLAY Yes See SQL/MX Report Writer Guide
Numeric data types Yes See SQL/MX Reference Manual
Numeric literals Yes See SQL/MX Reference Manual
OBEY command Yes See SQL/MX Reference Manual
OCTET_LENGTH function Yes See SQL/MX Reference Manual
OPEN statement Yes See SQL/MX Reference Manual
OSS names Yes
OUT command No

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2- 11
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 12 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
OUT_REPORT command Yes See SQL/MX Report Writer Guide
OVERFLOW_CHAR Yes See SQL/MX Report Writer Guide
command
OWNER file attribute No
PAGE_COUNT option Yes See SQL/MX Report Writer Guide
PAGE_FOOTING option Yes See SQL/MX Report Writer Guide
PAGE_LENGTH option Yes See SQL/MX Report Writer Guide
PAGE_NUMBER option Yes See SQL/MX Report Writer Guide
PAGE TITLE option Yes See SQL/MX Report Writer Guide
Parallel index loading Changed NonStop SQL/MX does parallel
index loading by default. It is not a
specifiable attribute.
Parameters Yes No TYPE AS clause. See SQL/MX
Reference Manual
PARTITION clause Yes See SQL/MX Reference Manual
Partitions Yes See SQL/MX Reference Manual
Partition management Yes Performed by SQL/MX MODIFY
command. See SQL/MX Reference
Manual
PARTNS table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
PERUSE command No
Plans Yes See SQL/MX Query Guide
POSITION function Yes See SQL/MX Reference Manual
Predicates Yes See Section 3, DML Features and
SQL/MX Reference Manual
PREPARE statement Yes See SQL/MX Reference Manual and
SQL/MX Programming Guide for C
and COBOL
Primary keys Yes See SQL/MX Reference Manual
Print item No
PROGID file attribute No

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2- 12
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 13 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
Program invalidation No
PROGRAMS table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
Protection view No See Views on page 6-6
PURGE command No
PURGEDATA command Yes See SQL/MX Reference Manual
Qualified fileset list No
Quantified predicate Yes See SQL/MX Reference Manual
RECLENGTH file attribute No
RELEASE statement No
REPORT FOOTING Yes See SQL/MX Report Writer Guide
command
REPORT option No
REPORT TITLE command Yes See SQL/MX Report Writer Guide
Reserved words Yes See SQL/MX Reference Manual
RESETBROKEN file attribute Changed Replaced by mxtool FIXUP
command
RESET DEFINE statement No See SQL/MX Reference Manual
RESET LAYOUT command Yes See SQL/MX Report Writer Guide
RESET PARAM command Yes See SQL/MX Reference Manual
RESET PREPARED No
command
RESET REPORT command Yes See SQL/MX Report Writer Guide
RESET SESSION command Yes See SQL/MX Report Writer Guide
RESET STYLE command Yes See SQL/MX Report Writer Guide
RESETBROKEN file attribute No
RIGHT_MARGIN option Yes See SQL/MX Report Writer Guide
ROLLBACK WORK Yes See SQL/MX Reference Manual
statement
ROWCOUNT option Yes See SQL/MX Report Writer Guide
Sample database Yes See SQL/MX Reference Manual

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2- 13
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 14 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
SAVE command No
Search conditions Yes See SQL/MX Reference Manual
SECURE command No Replaced by GRANT and REVOKE
statements. See SQL/MX Reference
Manual
SECURE file attribute No
Security Yes See SQL/MP Reference Manual
SELECT statement Yes Some differences. See SQL/MX
Reference Manual
SERIALWRITES file attribute No
SET DEFINE statement Yes See SQL/MX Reference Manual
SET DEFMODE command No
SET LAYOUT command Yes See SQL/MX Report Writer Guide
SET PARAM command Yes See SQL/MX Reference Manual
See SQL/MX Report Writer Guide
SET SESSION command No
SET STYLE command Yes See SQL/MX Report Writer Guide
SETSCALE function Changed See Section 4, Embedded SQL
Shorthand view No See Views on page 6-6
SHOW CONTROL command Yes See SQL/MX Reference Manual
SHOW DEFINE statement Yes See SQL/MX Reference Manual
SHOW DEFMODE No
command
SHOW LAYOUT command Yes See SQL/MX Report Writer Guide
SHOW PARAM command Yes See SQL/MX Reference Manual
SHOW PREPARED No
command
SHOW REPORT command Yes See SQL/MX Report Writer Guide
SHOW SESSION command Yes See SQL/MX Reference Manual
See SQL/MX Report Writer Guide
SHOW STYLE command Yes SQL/MX Report Writer Guide

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2- 14
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 15 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
Similarity checks Changed Handled internally by NonStop
SQL/MX. See Similarity Checks and
Automatic Recompilation on
page 4-13
SLACK file attribute No
SMF No SQL/MX objects cannot be on
SMF-managed volumes.
SPACE option Yes See SQL/MX Report Writer Guide
SQL directive Yes See SQL/MX Programming Guide
for C and COBOL
SQL identifiers Yes See SQL/MX Reference Manual
SQLDA area Changed See Dynamic SQL on page 4-6
SQLCI Changed MXCI in NonStop SQL/MX. See
SQL/MX Reference Manual for
MXCI command. See SQL/MP
Reference Manual for SQLCI
command.
SQLCI commands Changed MXCI in NonStop SQL/MX. See
MXCI Commands in SQL/MX
Reference Manual
SQLCODE Changed See SQL/MX Programming Guide
for C and COBOL
SQLCOMP command Changed MXCI in NonStop SQL/MX. See
SQL/MX Programming Guide for C
and COBOL
SQLSA area No See Statistics Area on page 4-10
Standards conformance Yes See Section 1, Comparing the
Differences in the Products and
SQL/MX Reference Manual
Statements Yes Some differences. See SQL DML
Statements on page 3-12 and
SQL/MX Reference Manual
Static SQL Yes See SQL/MX Programming Guide
for C and COBOL

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2- 15
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 16 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
Statistics Changed See Section 5, Optimizer and
Executor and SQL/MX Reference
Manual
HP NonStop Storage No SQL/MX objects cannot be on
Management Foundation SMF-managed volumes.
(SMF)
Stored procedures Yes See SQL/MX Guide to Stored
Procedures in Java
String functions Yes See Section 3, DML Features and
SQL/MX Reference Manual
String literals Yes See SQL/MX Reference Manual
Subqueries Yes See SQL/MX Reference Manual
SUBSTRING function Yes See SQL/MX Reference Manual
SUBTOTAL command Yes See SQL/MX Report Writer Guide
SUBTOTAL_LABEL option Yes See SQL/MX Report Writer Guide
SUM function Yes See SQL/MX Reference Manual
Super ID Yes See Authorization ID in SQL/MX
Reference Manual
SYSKEY Yes See SQL/MX Reference Manual
System catalog Changed Known as Metadata in NonStop
SQL/MX. See SQL/MX Reference
Manual
SYSTEM command No
System DEFINEs Yes See SQL/MX Reference Manual
TABLECODE file attribute No See SQL/MP Reference Manual
Tables Yes See SQL/MX Reference Manual
TABLES table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
TEDIT command No
Temporary tables Yes Used internally only in NonStop
SQL/MX.
TIME_FORMAT option Yes See SQL/MX Report Writer Guide

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2- 16
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 17 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
TIME data type Yes See Section 3, DML Features and
SQL/MX Reference Manual
TIMESTAMP data type Yes See Section 3, DML Features and
SQL/MX Reference Manual
TMF transactions Yes See Transaction Management
Facility (TMF) Introduction
TOTAL command No
TRANSID table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
TRIM function Yes See SQL/MX Reference Manual
UNDERLINE_CHAR option Yes See SQL/MX Report Writer Guide
UNLOCK TABLE statement Yes See SQL/MX Reference Manual
UPDATE statement Yes See SQL/MX Reference Manual and
SQL/MX Programming Guide for C
and COBOL
UPDATE STATISTICS Changed See Section 5, Optimizer and
statement Executor, SQL/MX Query Guide,
and SQL/MX Reference Manual
UPGRADE CATALOG No
command
UPGRADE SYSTEM No
CATALOG command
UPSHIFT function Yes See SQL/MX Reference Manual
USAGES table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
User-defined keys Yes See SQL/MX Reference Manual
Utilities Changed See Section 7, Utility Differences
VARCHAR_WIDTH option Yes See SQL/MX Report Writer Guide
VERIFIEDWRITES file No
attribute
VERIFY command Yes Replaced by mxtool VERIFY
command

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2- 17
Quick Reference to Differences

Table 2-1. NonStop SQL/MP Contents Compared to NonStop

SQL/MX (page 18 of 18)
Implemented
in NonStop
NonStop SQL/MP Item SQL/MX? Comments
Versions No
VERSIONS table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
Views Changed No protection or shorthand views.
See Views on page 6-6.
VIEWS table Changed Metadata tables are implemented
differently. See SQL/MX Reference
Manual
VOLUME command No
WHENEVER directive Yes See Section 4, Embedded SQL and
SQL/MX Programming Guide for C
and COBOL
WHERE clause Yes See SQL/MX Reference Manual
WINDOW option Yes See SQL/MX Report Writer Guide
WITH SHARED ACCESS No
option
! command Yes

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

2- 18
 3 DML Features
NonStop SQL/MX contains many SQL/MP extensions to the ANSI standard to enable
users to make a smooth transition to NonStop SQL/MX. See “Using NonStop SQL/MX
to Access SQL/MP Databases” in the SQL/MX Reference Manual if you have
questions about how to use SQL/MP tables with the SQL/MX compiler and executor.
Also, see the specific information you need to use date-time data types and literals,
INTERVAL data types and literals, and CHARACTER string data types and literals.
Although the current release of NonStop SQL/MP conforms to the SQL-89 standard,
some major differences exist between NonStop SQL/MP and SQL:1999 in the areas of
data types, expressions and functions, DML statement syntax, access options, and
transaction management. These areas also show up as differences between NonStop
SQL/MP and NonStop SQL/MX that are discussed in this section.

Data Types
NonStop SQL/MX supports most SQL/MP data types. Some SQL/MP DATETIME data
types are mapped to the SQL/MX SQL:1999 date-time data types DATE, TIME, and
TIMESTAMP. NonStop SQL/MX recognizes the nonstandard FRACTION keyword and
maps it directly to the fractional precision of the corresponding SQL:1999 compliant
data type.

Date-time Data Types

Both NonStop SQL/MP and NonStop SQL/MX support the SQL:1999 date-time data
types of DATE, TIME, and TIMESTAMP. Both NonStop SQL/MP and NonStop SQL/MX
support a number of nonstandard date-time types with specific start-end fields.
However, NonStop SQL/MX maps some date-time data types to DATE, TIME and
TIMESTAMP data types. NonStop SQL/MX does not support nonstandard date-time
data types with a start field of FRACTION.
Table 3-1 lists the mapping between date-time data types used in NonStop SQL/MP
and data types supported in NonStop SQL/MX. For information about how NonStop
SQL/MX processes nonstandard date-time data types, see the SQL/MX Reference
Manual.

Selecting Supported DATETIME Columns

NonStop SQL/MX supports accessing any SQL/MP DATETIME column except those
consisting of FRACTION only.
Suppose that an SQL/MP table has a DATETIME column defined as:
MPDateTimeCol DATETIME MONTH TO DAY
DEFAULT DATETIME '03-12' MONTH TO DAY
In NonStop SQL/MX you can select from this SQL/MP column as follows:

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3 -1
DML Features Interval Data Type

SELECT MPDateTimeCol FROM MPTable;

MPDateTimeCol
-------------
...
03-12
...

Selecting FRACTION-Only DATETIME Columns

If you attempt to select data from a FRACTION-only DATETIME column, the value is
returned as a string of '#' characters with the same display length as the length of the
column.
Suppose that an SQL/MP table has a DATETIME column defined as:
MPDateTimeCol DATETIME FRACTION(6)
DEFAULT DATETIME '123456' FRACTION(6)
You cannot select the data from this column. For example:
SELECT MPDateTimeCol FROM MPTable;
MPDateTimeCol
-------------
######
######
...
NonStop SQL/MX returns a warning indicating that you selected an unsupported data
type with undefined contents. However, if you try to use the unsupported data type in
any sort of expression, NonStop SQL/MX returns an error.
The SQL/MP DATETIME columns that do not map to standard SQL/MX types are
represented as SQL/MP DATETIME types in NonStop SQL/MX. These types are listed
in Table 3-1.
In embedded SQL, the method for declaring input and output host variables with
date-time data types for NonStop SQL/MX is different from the SQL/MP method. See
Section 4, Embedded SQL for more information about host variables.

Interval Data Type

NonStop SQL/MX supports the SQL:1999 INTERVAL data type. NonStop SQL/MX
recognizes the nonstandard FRACTION keyword used in NonStop SQL/MP and maps
it to the equivalent fractional portion of the standard SECOND field. However, it does
not support interval data types with a start field of FRACTION.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3 -2
DML Features Selecting FRACTION-Only Interval Columns

This table lists the mapping between the interval data types used in NonStop SQL/MP
and those used in NonStop SQL/MX:

NonStop SQL/MP Interval Type Mapping in NonStop SQL/MX

INTERVAL DAY(n) TO SECOND INTERVAL DAY(n) TO SECOND(0)
INTERVAL DAY(n) TO FRACTION(f) INTERVAL DAY(n) TO SECOND(f)
INTERVAL HOUR(n) TO SECOND INTERVAL HOUR(n) TO SECOND(0)
INTERVAL HOUR(n) TO FRACTION(f) INTERVAL HOUR(n) TO SECOND(f)
INTERVAL MINUTE(n) TO SECOND INTERVAL MINUTE(n) TO SECOND(0)
INTERVAL MINUTE(n) TO FRACTION(f) INTERVAL MINUTE(n) TO SECOND(f)
INTERVAL SECOND(n) TO SECOND INTERVAL SECOND(n,0)
INTERVAL SECOND(n) TO FRACTION(f) INTERVAL SECOND(n,f)
INTERVAL FRACTION(n) TO FRACTION(f) Not supported

NonStop SQL/MX handles the interval data types in the next list in the same manner
as NonStop SQL/MP:
• YEAR TO (YEAR, MONTH)
• MONTH TO MONTH
• DAY TO (DAY, HOUR, MINUTE)
• HOUR TO (HOUR, MINUTE)
• MINUTE TO MINUTE
For precision and default of interval data types, NonStop SQL/MP and NonStop
SQL/MX are the same. For additional information about interval data types, see the
SQL/MX Reference Manual.

Selecting FRACTION-Only Interval Columns

If you attempt to select data from a FRACTION-only interval column, the value is
returned as a string of '#' characters with the same display length as the length of the
column.
Suppose that an SQL/MP table has an INTERVAL column defined as:
MPIntervalCol INTERVAL FRACTION(6)
DEFAULT INTERVAL '123456' FRACTION(6)
You cannot select the data from this column. For example, if you enter this statement:
SELECT MPIntervalCol FROM MPTable;
MPIntervalCol
-------------
######
######
...

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3 -3
DML Features Floating-Point Data Types

NonStop SQL/MX returns a warning indicating that you selected an unsupported data
type with undefined contents. If you try to use the unsupported data type in any sort of
expression, NonStop SQL/MX returns an error.

Floating-Point Data Types

NonStop SQL/MP supports floating-point data types of FLOAT, REAL, and DOUBLE
PRECISION. A column in an SQL/MP table declared with a floating-point data type is
stored in Tandem floating-point format, and all computations on it are done assuming
that.
In previous releases of NonStop SQL/MX, if you want an application to input or output
a date-time or interval value, you must cast it to STRING data type and then input or
output it to or from NonStop SQL/MP.This makes the use of date-time or interval data
types cumbersome in application programs and is not ANSI compliant.
SQL/MX tables now store columns in IEEE floating-point format, as data types of
FLOAT, REAL, and DOUBLE PRECISION. This allows input and output of data from
and to application host variables in Tandem or IEEE floating-point.
NonStop SQL/MX can select data from both SQL/MX and SQL/MP tables. The
returned data from mxci is always in IEEE format. The returned data in an embedded
program can be in IEEE or Tandem format.
NonStop SQL/MX does not allow any column of the partitioning key of a partitionable
table or index to be float data type. If there is no partitioning key specified, the
clustering key becomes the partitioning key.
For information about how SQL/MX floating-point data types, see “Numeric Data
Types” in the SQL/MX Reference Manual.

File Types
SQL/MP tables can have one of three physical file organizations: key-sequenced,
entry-sequenced, or relative and can be partitioned. You can access these types of
SQL/MP files through NonStop SQL/MX:
• Key-sequenced tables with or without partitions
• Entry-sequenced tables that are not partitioned
You cannot access these types of SQL/MP files through NonStop SQL/MX:
• Entry-sequenced tables that are partitioned
• Relative tables
For more information about SQL/MP file organizations, see the SQL/MP Reference
Manual.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3 -4
DML Features Identifiers

Identifiers
NonStop SQL/MP supports only regular identifiers, formed from alphanumeric
characters and the underscore character. Regular identifiers are not case sensitive. In
NonStop SQL/MP, an SQL identifier can contain up to 30 letters (A through Z or a
through z), digits (0 through 9), or underscore (_) characters. The first character must
be a letter.
NonStop SQL/MX supports both regular and delimited identifiers. An identifier of either
type can contain up to 128 characters. Regular identifiers are defined as they are in
NonStop SQL/MP—with the exception of length. Delimited identifiers are character
strings that appear within double quote characters (") and consist of alphanumeric
characters and any character in the SQL-92 special character set except for the at sign
(@), the forward slash (/), backward slash (\), and circumflex (^).
For additional information about identifiers, see the SQL/MX Reference Manual.

Reserved Words
In NonStop SQL/MP, you should not use reserved words as names of constraints,
columns (including correlation names), cursors, or statements. You can, however, use
reserved words in catalog names, in Guardian names that identify tables, indexes,
views, partitions, collations, and in host variables.
NonStop SQL/MX reserves a longer list of words and allows the use of a reserved
word only if you place double quotes around it and observe case-sensitivity. You
cannot access views that are defined on tables named after reserved words. See the
SQL/MP Reference Manual for a list of SQL/MP reserved words. See the SQL/MX
Reference Manual for a list of SQL/MX reserved words. In addition, see the Identifiers
entry in the SQL/MX Reference Manual for information about how reserved words can
be used and limitations in NonStop SQL/MX.

Literals
NonStop SQL/MX uses single quote characters (') as character string delimiters.
NonStop SQL/MX supports all SQL/MP date-time literals. However, if you try to use a
literal with unsupported FRACTION-only INTERVAL or DATETIME in any sort of
expression, NonStop SQL/MX returns an error.
For additional information about literals, see the SQL/MX Reference Manual.

String Delimiters
NonStop SQL/MP uses both single (') and double quote characters (") to enclose
character, date-time, and interval literals. NonStop SQL/MX uses only single quote
characters for string delimiters, conforming to the SQL:1999 standard. NonStop
SQL/MX uses double quotes for delimited identifiers.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3 -5
DML Features Date-time Literals

Date-time Literals
In NonStop SQL/MP, a date-time literal is a DATETIME, DATE, TIME, or TIMESTAMP
literal that you can use in an expression and which can appear in default, USA, or
European format.
In NonStop SQL/MX, a date-time literal is a DATE, TIME, or TIMESTAMP literal. An
SQL/MX date-time literal begins with the DATE, TIME, or TIMESTAMP keyword and
can appear in default, USA, or European format.
For example, these DATE literals are in default, USA, and European formats,
respectively:
DATE '1990-01-22'
DATE '01/22/1990'
DATE '22.01.1990'

Supported and Unsupported SQL/MP Date-Time Literals in

NonStop SQL/MX
NonStop SQL/MP date-time literals fall into one of these categories:
• Those that map directly to SQL:1999 types and are supported by NonStop
SQL/MX
• Those that do not map to SQL:1999 types but are supported by NonStop SQL/MX
• Those that do not map to SQL:1999 types and are not supported in NonStop
SQL/MX
NonStop SQL/MP date-time literals that are not supported are listed in Table 3-1:
:

Table 3-1. Support For NonStop SQL/MP DATETIME Literals in NonStop

SQL/MX (page 1 of 3)
Supported
in NonStop
NonStop SQL/MP DATETIME Type SQL/MX Maps To ANSI Type
DATETIME YEAR TO DAY Yes DATE
DATETIME YEAR TO SECOND Yes TIMESTAMP(0)
DATETIME YEAR TO FRACTION (1) Yes TIMESTAMP(1)
DATETIME YEAR TO FRACTION (2) Yes TIMESTAMP(2)
DATETIME YEAR TO FRACTION (3) Yes TIMESTAMP(3)
DATETIME YEAR TO FRACTION (4) Yes TIMESTAMP(4)
DATETIME YEAR TO FRACTION (5) Yes TIMESTAMP(5)
DATETIME YEAR TO FRACTION (6) Yes TIMESTAMP(6)
DATETIME HOUR TO SECOND Yes TIME(0)
DATETIME HOUR TO FRACTION (1) Yes TIME(1)
DATETIME HOUR TO FRACTION (2) Yes TIME(2)

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3 -6
DML Features Date-time Literals

Table 3-1. Support For NonStop SQL/MP DATETIME Literals in NonStop

SQL/MX (page 2 of 3)
Supported
in NonStop
NonStop SQL/MP DATETIME Type SQL/MX Maps To ANSI Type
DATETIME HOUR TO FRACTION (3) Yes TIME(3)
DATETIME HOUR TO FRACTION (4) Yes TIME(4)
DATETIME HOUR TO FRACTION (5) Yes TIME(5)
DATETIME HOUR TO FRACTION (6) Yes TIME(6)
DATETIME YEAR Yes (None)
DATETIME YEAR TO MONTH Yes (None)
DATETIME YEAR TO HOUR Yes (None)
DATETIME YEAR TO MINUTE Yes (None)
DATETIME MONTH Yes (None)
DATETIME MONTH TO DAY Yes (None)
DATETIME MONTH TO HOUR Yes (None)
DATETIME MONTH TO MINUTE Yes (None)
DATETIME MONTH TO SECOND Yes (None)
DATETIME MONTH TO FRACTION(n) Yes (None)
DATETIME DAY Yes (None)
DATETIME DAY TO HOUR Yes (None)
DATETIME DAY TO MINUTE Yes (None)
DATETIME DAY TO SECOND Yes (None)
DATETIME DAY TO FRACTION(n) Yes (None)
DATETIME HOUR Yes (None)
DATETIME HOUR TO MINUTE Yes (None)
DATETIME HOUR TO SECOND Yes (None)
DATETIME HOUR TO FRACTION(n) Yes (None)
DATETIME MINUTE Yes (None)
DATETIME MINUTE TO SECOND Yes (None)
DATETIME MINUTE TO FRACTION(n) Yes (None)
DATETIME SECOND Yes (None)
DATETIME SECOND TO FRACTION(n) Yes (None)
DATETIME FRACTION TO FRACTION(n) No* (None)
INTERVAL SECOND(0) TO FRACTION (n) No* (None)
INTERVAL FRACTION TO FRACTION (n) No* (None)

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3 -7
DML Features Interval Literals

Table 3-1. Support For NonStop SQL/MP DATETIME Literals in NonStop

SQL/MX (page 3 of 3)
Supported
in NonStop
NonStop SQL/MP DATETIME Type SQL/MX Maps To ANSI Type
INTERVAL start to FRACTION(n) Yes INTERVAL start to
SECOND(n)
All other INTERVAL start-end combinations Yes Same as in NonStop
SQL/MP
*Limited support as dummy values in SELECT
statements only

Interval Literals
As in NonStop SQL/MP, an SQL/MX interval literal is either year-month or day-time. If
you are using SQL/MX DML statements with SQL/MP tables, NonStop SQL/MX
recognizes fields created with the FRACTION keyword and maps them directly to the
fractional precision of the SECOND field.
These are INTERVAL literals:
INTERVAL '7' DAY
INTERVAL '2-7' YEAR TO MONTH
INTERVAL '5:2:15:36.33' DAY TO SECOND(2)
An SQL/MP interval with a literal of DAY TO FRACTION(2) becomes in SQL/MX DAY
TO SECOND(2).

Expressions and Functions

NonStop SQL/MP supports three types of expressions: value expressions, row-value
constructors, and query expressions. NonStop SQL/MX supports SQL:1999 versions of
value expressions, row-value constructors, and query expressions. For a complete list
of SQL/MX expressions and functions, see the SQL/MX Reference Manual.
Table 3-2 lists only the SQL/MP functions that are obsolete or changed in NonStop
SQL/MX:

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3 -8
DML Features Numeric Value Expressions and Functions

Table 3-2. NonStop SQL/MP and NonStop SQL/MX Functions

Implemented
NonStop SQL/MP in NonStop
Function SQL/MX? Comments
COMPUTE_TIMESTAMP No
CURRENT Changed In NonStop SQL/MX, supported to return a
range of date-time values only in NonStop
SQL/MP stored text. Otherwise, in NonStop
SQL/MX, same as CURRENT_TIMESTAMP.
See mapping under Datetime Functions.
EXTEND Changed In NonStop SQL/MX, supported only for use
in NonStop SQL/MP stored text. Otherwise,
use the CAST specification in NonStop
SQL/MX.
SETSCALE Changed In NonStop SQL/MX, you provide the scale
when you declare your host variables in your
program. See SETSCALE Function.

Except for the functions listed previously, NonStop SQL/MX offers all the functions of
NonStop SQL/MP and adds many new functions.

Numeric Value Expressions and Functions

Both NonStop SQL/MX and NonStop SQL/MP offer an extension to SQL:1999, the
exponentiation operator (**), but in results of computations of numeric expressions,
NonStop SQL/MX might differ in precision and scale from the computations within
NonStop SQL/MP.

Datetime Functions
In NonStop SQL/MP, the CURRENT function returns a current local DATETIME value,
based on the specified start-end fields, such as CURRENT YEAR TO MONTH,
CURRENT YEAR or CURRENT SECOND. CURRENT by itself is the equivalent of
CURRENT YEAR TO FRACTION.
For the same results in NonStop SQL/MX, use the CURRENT_DATE,
CURRENT_TIME, or CURRENT_TIMESTAMP function. In addition, CURRENT is an
SQL/MX extension that maps to CURRENT_TIMESTAMP, equivalent to its result in
NonStop SQL/MP. NonStop SQL/MX does not support the use of CURRENT with start-
end fields. For this result, you should cast the appropriate form of SQL/MX
CURRENT_DATE, CURRENT_TIME, or CURRENT_TIMESTAMP to the SQL/MP
DATETIME type with the desired start-end fields. Table 3-3 compares SQL/MP and
SQL/MX date-time functions:

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3 -9
DML Features Interval Value Expressions

Table 3-3. Mapping NonStop SQL/MP Datetime Functions To NonStop SQL/MX

NonStop SQL/MP NonStop SQL/MX
CURRENT CURRENT_TIMESTAMP
CURRENT YEAR TO DAY CURRENT_DATE
CURRENT YEAR TO SECOND CURRENT_TIMESTAMP(0)
CURRENT YEAR TO FRACTION(n) CURRENT_TIMESTAMP(n)
CURRENT HOUR TO SECOND CURRENT_TIME(0)
CURRENT HOUR TO FRACTION(n) CURRENT_TIME(n)
CURRENT YEAR CAST (CURRENT_DATE AS DATETIME YEAR)
CURRENT YEAR TO MONTH CAST (CURRENT_DATE AS DATETIME YEAR
TO MONTH)
CURRENT HOUR TO MINUTE CAST (CURRENT_TIME AS DATETIME HOUR
TO MINUTE)

Interval Value Expressions

NonStop SQL/MP allows an interval to be divided by an interval. For example,
INTERVAL '6' DAY / INTERVAL '2' DAY results in an integer value of 3. NonStop
SQL/MP does not support the CAST function for date-time and interval data types.
NonStop SQL/MX (and SQL:1999) supports the division (or multiplication) of an
INTERVAL by a factor. For example, INTERVAL '6' DAY / 2 results in an interval of 3
days. If you want the result to be an integer and not an interval, convert or cast the
result as an integer.

Mathematical Functions
NonStop SQL/MP does not support mathematical functions. NonStop SQL/MX
supports many mathematical functions as extensions to the SQL-92 standard.

Sequence Functions
NonStop SQL/MP does not support sequence functions. In NonStop SQL/MX,
sequence functions operate on ordered rows of a SELECT result table that includes a
SEQUENCE BY clause.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 10
DML Features Aggregate Functions

Aggregate Functions
In a few cases, NonStop SQL/MP allows a subquery within an aggregate in a
predicate. NonStop SQL/MX follows the ANSI standard and does not allow any
subqueries in an aggregate.
In order to comply with ANSI standards, NonStop SQL/MX does not move aggregate
predicates from the WHERE clause to a HAVING clause and does not move non-
aggregate predicates from the HAVING clause to the WHERE clause, as NonStop
SQL/MP does.

Predicates
Both NonStop SQL/MP and NonStop SQL/MX support the use of row-value
constructors within predicates. However, the NonStop SQL/MP implementation is not
SQL:1999 compliant. In NonStop SQL/MP, you are not required to enclose the
sequence of expressions of a row-value constructor in parentheses. In NonStop
SQL/MX, it is helpful (although not required) to do so. If you do not enclose the
sequence of expressions in parentheses, your resulting queries might be ambiguous.
For example, in this predicate, consisting of a comparison predicate and a NULL
predicate, SQL/MP row-value constructors use no parentheses:
WHERE a,b > 10,c AND p,q,r IS NULL
In contrast, in NonStop SQL/MX, the row-value constructors in this predicate use
parentheses:
WHERE (a,b) > (10,c) AND (p,q,r) IS NULL
Use row-value constructors in comparison predicates, the IN predicate, the NULL
predicate, and the quantified comparison predicates (ALL, ANY, SOME).

LIKE Predicate
Both NonStop SQL/MP and NonStop SQL/MX support the LIKE predicate. However,
NonStop SQL/MP has an extension to the ANSI standard. It allows you to specify a
TERMINATE character to indicate the end of the pattern within the pattern string. You
can use this clause when the column value and the comparison value are of different
lengths. NonStop SQL/MX supports the SUBSTRING and POSITION functions, which
can be used for the same purpose as TERMINATE.
NonStop SQL/MX rewrites SQL/MP constraint and view text that uses the LIKE
predicate and TERMINATE character:

Original SQL/MP predicate

column1 LIKE ?expr2 TERMINATE ?trm
column1 LIKE ?expr2 ESCAPE ?esc TERMINATE ?trm

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 11
DML Features Sort Operations

Rewritten by NonStop SQL/MX

column1 LIKE SUBSTRING(?expr2 FROM 1
FOR (POSITION (?trm in ?expr2) - 1))
column1 LIKE SUBSTRNG(?expr2 FROM 1
FOR (POSITION(?trm IN ?expr2) - 1)) ESCAPE ?esc

Sort Operations
NonStop SQL/MP uses the FastSort sort/merge program for certain operations. These
operations are described in the SQL/MP Installation and Management Guide. You use
the EXPLAIN utility to determine if a sort occurs. The EXPLAIN reports any sort
operations required to execute the query. You also use certain SORT-related DEFINEs
to influence the effectiveness of some SQL operations.
NonStop SQL/MX does not use the FastSort program or the SORT-related DEFINEs.
In addition, a scratch file in NonStop SQL/MX can overflow to another disk. The
operation will not fail if a scratch file becomes full or if the disk becomes full; NonStop
SQL/MX creates an additional scratch file either on the same or a different disk if the
disk becomes full. The maximum number of scratch files that can be used is 2048.
NonStop SQL/MX provides default settings that influence the choice of scratch disks.
SQL/MX parallelism is performed through ESPs only. Parallel sort can be forced
through CONTROL QUERY SHAPE. When the SQL/MX optimizer runs parallel
queries with ESPs, each ESP tries to use a different disk.
For additional information about how scratch disks are selected and influenced, see
the Defaults section of the SQL/MX Query Guide and the SQL/MX Reference Manual.

SQL DML Statements

Most SQL/MP DML statements exist in NonStop SQL/MX. However, a few options and
clauses have been dropped because of SQL:1999 ANSI changes. Table 3-4 lists the
SQL/MP DML statements that differ in NonStop SQL/MX:

Table 3-4. SQL/MX and SQL/MP DML Statements

Implemented
SQL/MP DML in NonStop
Statements SQL/MX? Comments
CONTROL EXECUTOR No PARALLEL EXECUTION is ON by default in
NonStop SQL/MX. See Default Settings on
page 5-7.
CONTROL QUERY Changed See CONTROL QUERY Directive on page 3-23.
CONTROL TABLE Changed See CONTROL TABLE Directive on page 3-21;
also see Forcing Query Execution Plans on
page 5-8.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 12
DML Features Nonaudited Tables

Table 3-4. SQL/MX and SQL/MP DML Statements

Implemented
SQL/MP DML in NonStop
Statements SQL/MX? Comments
INSERT Changed In NonStop SQL/MX, an ORDER BY clause can
be used following the specification of the insert
source. Also, multiple row-value constructors can
be used following the VALUES keyword. See
INSERT Statement on page 3-14.
SELECT Changed In NonStop SQL/MX, an SQL value expression
can be used in a select list. Also, derived tables
(or query expressions) can be used in a FROM
clause. See SELECT Statement on page 3-15.
UPDATE Changed In NonStop SQL/MX, a scalar subquery can be
used as the update value in a SET clause. See
UPDATE Statement on page 3-16.

Nonaudited Tables
In both NonStop SQL/MP and NonStop SQL/MX, the HP NonStop Transaction
Management Facility (TMF) product works only on audited tables, so a transaction
does not protect operations on nonaudited tables. The simplest approach is to make all
tables audited.
NonStop SQL/MP is designed to avoid any inconsistencies between the base table and
an index for nonaudited tables. Because of its design, NonStop SQL/MX has a greater
likelihood that such inconsistencies might occur.
Nonaudited tables are not protected by transactions and follow a different locking and
error handling model than audited tables. In NonStop SQL/MX, certain situations such
as DML error occurrences or utility operations with DML operations can lead to
inconsistent data within a nonaudited table or between a nonaudited table and its
indexes.
To avoid potential problems, do not create indexes on nonaudited tables. Do not run
DDL or utility operations concurrently with DML operations on nonaudited tables. This
table lists differences in how NonStop SQL/MP and NonStop SQL/MX handle
nonaudited DML operations
Operation NonStop SQL/MP NonStop SQL/MX
Insert/update/delete with a user error No inconsistencies Potentially many rows
during index maintenance affected
Insert/update/delete interrupted by At most one row Potentially many rows
internal program abend or CPU halt affected

NonStop SQL/MX provides a default setting that can warn you at compile time of
nonaudited index maintenance, called IUD_NONAUDITED_INDEX_MAINT. By default,
operations that could lead to an inconsistency between base tables and indexes are

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 13
DML Features INSERT Statement

disabled. See the Defaults Table entry in the SQL/MX Reference Manual for further
information.

INSERT Statement

RETURNING LASTSYSKEY Clause

NonStop SQL/MP supports the RETURNING LASTSYSKEY clause for the INSERT
statement. NonStop SQL/MX does not support this option.

Insert Rows in Order

NonStop SQL/MX allows you to insert rows in order with an ORDER BY clause on the
INSERT statement. When selecting rows from one table to be inserted into another, an
ORDER BY clause can reorder the sequence of the insertion so that the inserts are
done more efficiently. You cannot use ORDER BY in NonStop SQL/MX when the
SELECT statement is a subquery. The ORDER BY clause cannot be used in
embedded SQL programs. The ORDER BY clause is an SQL/MX extension to the
ANSI-92 standard. For additional information about the ORDER BY clause, see the
SQL/MX Reference Manual.

Inserting Multiple Rows

In NonStop SQL/MP, if you use the VALUES keyword, you can insert only one row. For
example, this SQL/MP statement inserts a single row of values:
INSERT INTO JOB VALUES (1000, 'MANAGER');
In NonStop SQL/MX, if you use the VALUES keyword, you can insert multiple row-
value constructors. For example, this SQL/MX statement inserts two rows into table
JOB1:
INSERT INTO JOB1 VALUES (1000, 'MANAGER'), (2000, 'SUPERVISOR')
In general, in NonStop SQL/MX, each row to be inserted is specified as a list of value
expressions or a row subquery. A value in a row can also be a scalar subquery. For
example, this SQL/MX statement inserts two rows into table JOB:
INSERT INTO JOB VALUES (1000, 'MANAGER'), ((SELECT *
FROM JOB1 WHERE JOBCODE = 2000));
In NonStop SQL/MP you usually perform multiple row inserts by setting a single row
insert to run in a loop. Even though NonStop SQL/MP is inserting a single row,
because it is looped, it will do this insert as a multiple row insert in DP2 using virtual
sequential block buffering (VSBB).
NonStop SQL/MX’s architecture handles this differently and it is more efficient to use
rowsets to perform multiple row inserts. If VSBB is on, EXPLAIN output will show
INSERT_VSBB as the operator.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 14
DML Features SELECT Statement

SELECT Statement

Table Reference Within FROM Clause

In NonStop SQL/MP, a table reference within a FROM clause of a SELECT statement
can be a table, view, or joined table. The maximum number of tables that can be
specified in a FROM clause is 16. This maximum includes the underlying base tables
of views.
In NonStop SQL/MX, a table reference can also be a query expression or, stated more
precisely, the derived table determined by the evaluation of a query expression. A
query expression can be the UNION or JOIN of two (or more) table references. A
query expression can also be a simple table—a table value constructor generated by a
VALUES clause, TABLE clause, or a subquery.
In NonStop SQL/MX, an item in the select list can be a subquery:
SELECT (SELECT a FROM s WHERE b=t.b) FROM t;
In NonStop SQL/MX, a joined table can be used as a table reference in the FROM
clause of a SELECT statement. An SQL:1999 joined table is specified by connecting
the names of the two tables with the keywords NATURAL, CROSS, INNER, LEFT, or
RIGHT preceding the JOIN keyword. NonStop SQL/MP supports LEFT and INNER
joins only.
For example, the clause PARTS NATURAL JOIN ODETAIL (as a table reference in a
FROM clause) joins rows only where the values of all columns that have the same
name in both tables match. The clause PARTS NATURAL LEFT JOIN ODETAIL joins
rows where the values of all columns that have the same name in both tables match
plus rows from PARTS (the table to the left of the JOIN keyword) that do not meet this
condition.

GROUP BY Clause and Numerical Arguments

In NonStop SQL/MP, the GROUP BY clause allows you to use numerical arguments,
for example:
SELECT a + b FROM t GROUP BY 1;
NonStop SQL/MX does not allow the use of numerical arguments with the GROUP BY
clause. However, you can use a derived query to achieve the same purpose as shown
next:
SELECT c FROM (SELECT a + b FROM t) x(c) GROUP BY c;

Placement of GROUP BY Clause

NonStop SQL/MX requires the GROUP BY clause, if one exists, to precede a HAVING
clause. In NonStop SQL/MP, you can place the clauses in any order.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 15
DML Features UPDATE Statement

UPDATE Statement

Source Value for a Column Referenced in the SET Clause

In NonStop SQL/MX, the source value for a column referenced in the SET clause of an
UPDATE statement is an SQL value expression. In particular, a value expression can
be a scalar subquery:
UPDATE JOB SET jobdesc=(SELECT jobdesc from JOB1
WHERE jobcode=2000)
WHERE jobcode=2000;
In NonStop SQL/MP, the expression cannot be a subquery.

Parallel Plans
In NonStop SQL/MP, you cannot have a parallel plan where the UPDATE statement
contains the WHERE CURRENT OF clause. In NonStop SQL/MX, you can have
parallel plans with the WHERE CURRENT OF clause.

Access Options and Isolation Levels

NonStop SQL/MP allows you to specify access options for each DML statement. In
NonStop SQL/MP, the statement level access options are BROWSE, STABLE, and
REPEATABLE.
NonStop SQL/MX allows you to specify the isolation level at both the statement and
transaction level. (Isolation level, SQL:1999 terminology, means the same as access
option.) In NonStop SQL/MX, the isolation levels for both statements and transactions
are READ UNCOMMITTED, READ COMMITTED, SERIALIZABLE (or REPEATABLE
READ) and STABLE. Isolation levels set at the statement level take precedence over
the transaction level settings. Isolation levels for statements are extensions to the ANSI
standard. Isolation levels for transactions are SQL:1999 compliant; you set transaction
isolation levels with the SET TRANSACTION statement.

Note. Unless you are writing applications for portability, you should use isolation levels at the
statement level. The SET TRANSACTION statement might cause a dynamic recompilation of
DML statements within the next transaction. Further, NonStop SQL/MX does not have the
opportunity to optimize query execution plans as it does during static compilation. You can
achieve better performance if you specify isolation levels for individual DML statements.

Isolation Levels Contrasted Between NonStop SQL/MP and

NonStop SQL/MX
• BROWSE (NonStop SQL/MP) and READ UNCOMMITTED (NonStop SQL/MX)
have the same implementation. This option is available only for the SELECT
statement.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 16
DML Features STABLE Access

• STABLE (NonStop SQL/MP) and STABLE (NonStop SQL/MX) have the same
implementation. This option is available only for the SELECT statement.
• REPEATABLE (NonStop SQL/MP and NonStop SQL/MX) and SERIALIZABLE
(NonStop SQL/MX) have the same implementation; this option holds locks on data
until the end of the transaction.

Note. If your SQL/MP application uses the BROWSE, STABLE, and REPEATABLE keywords,
NonStop SQL/MX accepts the keywords as synonyms for READ UNCOMMITTED, STABLE,
and SERIALIZABLE.

STABLE Access

How STABLE Access Works

In both SQL/MX and SQL/MP cursor operations using STABLE access, you are
guaranteed another transaction cannot update the row and that the row can be safely
updated without having to check versions of the row data (using UPDATE WHERE
CURRENT OF or DELETE WHERE CURRENT OF). The number of rows that are
locked depends on the value you have set for SYSTEM_DEFAULT attribute
MAX_ROWS_LOCKED_FOR_STABLE_ACCESS.
If you specify STABLE access for an updatable SELECT statement, NonStop SQL/MX
uses that access method. For other statements, NonStop SQL/MX changes the access
method to READ COMMITTED. NonStop SQL/MX does not choose STABLE access
for rowsets.

How READ COMMITTED Differs From STABLE Access

If you choose READ COMMITTED, read a row, and decide to update that row, in the
time between reading the row and making the update, the value might change because
no lock is acquired on the row. If your applications use STABLE access, you might
want to evaluate that use to decide whether you should change the access to
SERIALIZABLE or whether you should adjust your error handling procedures. You can
rely on the fact that when you read a row, no other transaction can change the row
while you are processing it. Thus, you are guaranteed not to cause any integrity issues
when you do update the row.
READ COMMITTED also guarantees integrity but in a different way. If you update the
row acquired using READ COMMITTED access and another transaction has updated it
since you first read it, a warning is returned to the application on the update. You can
check the warning code returned and take appropriate action (abort the transaction if
necessary, display an appropriate message, retry the read and update, and so on). By
changing your error handling procedures, you can benefit from the increased
concurrency.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 17
DML Features Transaction Management

STABLE Access and EXCLUSIVE MODE

In both NonStop SQL/MX and NonStop SQL/MP, if you use STABLE access with
EXCLUSIVE MODE, you do not get an error, because the next transaction must wait
for access to the row.

Transaction Management
A TMF transaction is the basic recoverable unit in case of a failure or transaction
interruption. TMF transactions can be defined during an MXCI session or in a host
program. The typical order of events is:
1. Transaction is started.
2. Database changes are made.
3. Transaction is committed.
If the changes cannot be made or you do not want to complete the transaction, your
program can abort the transaction so the database is rolled back to its original state.

Starting and Ending a Transaction

In NonStop SQL/MX and NonStop SQL/MP, to ensure that a sequence of statements
either executes successfully or not at all, define one transaction consisting of these
statements by using the BEGIN WORK and COMMIT WORK statements. You can
abort a transaction with the ROLLBACK WORK statement. BEGIN WORK is an
extension to the ANSI standard.

User-Defined or Explicit Transactions

Transactions you do not start through SQL, such as TMF, MXCS, or inherited
transactions, are called user-defined, or explicit, transactions. You must commit or roll
back these transactions.
If an error occurs in an explicit transaction for an INSERT, UPDATE, or DELETE
statement on audited tables, the transaction is not aborted automatically. However,
your data can be in an inconsistent state within a query.

System-Defined or Implicit Transactions

NonStop SQL/MP has implicit transactions in SQLCI but not in embedded statements.
NonStop SQL/MX also has implicit transactions. It initiates a transaction when it
encounters an SQL statement if a BEGIN WORK statement has not been specified
and if the statement references an audited table. Most DML statements initiate
transactions implicitly at the start of execution. NonStop SQL/MX does not initiate a
transaction for a SELECT statement with READ UNCOMMITTED access.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 18
DML Features SET TRANSACTION Statement

Even if a transaction is initiated implicitly, you must end a transaction explicitly with the
COMMIT WORK statement or the ROLLBACK WORK statement so that either the
entire SQL statement executes or none of it is does.

Auto-Abort
NonStop SQL/MX Release 1.8 automatically aborted transactions if an error occurred
while performing a SQL DELETE, INSERT, UPDATE, or DDL statement. By default,
NonStop SQL/MX Release 2.0 does not automatically abort transactions following an
error. The option to abort the transaction following an error must be determined by the
application program.
You can set a CONTROL QUERY DEFAULT, UPD_ABORT_ON_ERROR, to make
SQL/MX Release 2.0 behave like SQL/MX Release 1.8 and automatically abort
transactions following an error. See the description of this control query default in the
SQL/MX Reference Manual for details.

SQL/MP Behavior
In NonStop SQL/MP, you can use row-not-found warnings or row-already-exists errors
to determine whether to insert or update single rows. You can do an INSERT to find out
if the row already exists, ignore the error, and follow it with an UPDATE statement to
update the row. Or, do an UPDATE on a row to see if it exists, and when no row is
found, use an INSERT statement to insert the row. This process is sometimes called
an “UPSERT.”
SQL/MX Release 2.0 enables you to use row-not-found and row-already exists errors
the same way, by leaving the CONTROL QUERY DEFAULT
UPD_ABORT_ON_ERROR set to OFF, the default. See the description of this control
query default in the SQL/MX Reference Manual for details.

Inherited Transactions
In NonStop SQL/MP and NonStop SQL/MX, if a transaction is started at some point
during the execution of your program (for example, in TMF or MXCS), you do not need
to explicitly start a new SQL transaction. SQL detects that a transaction exists and
inherits that transaction. If an error occurs in an inherited transaction, NonStop
SQL/MX changes the transaction state to doomed (SQLSTATE “25000”). A doomed
transaction is not aborted. However, to proceed, you must manually abort the
transaction. NonStop SQL/MX provides error messages until then.

SET TRANSACTION Statement

In NonStop SQL/MX, to automatically end a transaction after a statement successfully
executes, use the AUTOCOMMIT option within a SET TRANSACTION statement in
both MXCI and embedded SQL. This option specifies whether or not NonStop SQL/MX
commits at the end of statement execution and applies to any statement for which SQL
initiates a transaction (implicit transaction). This option does not apply to transactions
you start with the BEGIN WORK statement or transactions started by services, such as

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 19
DML Features SET TRANSACTION Statement

TMF or MXCS, or inherited from other processes. NonStop SQL/MP provides a similar
SQLCI command, called AUTOWORK.
See Cursor Operations (Embedded SQL) on page 3-21 for information about using the
AUTOCOMMIT option with cursors.
The next statement sets the AUTOCOMMIT option to ON:
SET TRANSACTION AUTOCOMMIT ON;
If this option is set to ON, at the end of each statement execution, NonStop SQL/MX
automatically commits changes made to the database. AUTOCOMMIT is set to ON by
default at the start of an MXCI session.
If this option is set to OFF, the current transaction remains active until the end of the
MXCI session or, for embedded SQL, the end of program execution, at which time the
transaction is aborted—unless you explicitly commit or roll back the transaction. The
default is OFF for embedded SQL.
You can use the MXCI command ENV to check the state of the transaction.

AUTOCOMMIT Example
In the next example, the SET TRANSACTION statement sets AUTOCOMMIT to ON.
The first INSERT statement starts an implicit transaction. When the INSERT finishes
executing, the transaction, consisting of one inserted row, is committed. The next
INSERT statement occurs, and after it finishes, its transaction is committed. You start a
user-defined (explicit) transaction with the BEGIN WORK statement. A series of
UPDATE statements are contained in this user-defined transaction, and after the
COMMIT WORK statement is issued, the updates are committed. Remember that the
AUTOCOMMIT option is still set to ON, so that when the next DML statement is
initiated, an implicit transaction is started, and when the statement successfully
finishes, the transaction is committed.

SET TRANSACTION AUTOCOMMIT ON;

INSERT ...; --Implicit transaction started
--Row committed
INSERT ...; --Implicit transaction started
--Row committed
--AUTOCOMMIT STILL ON

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 20
DML Features CONTROL TABLE Directive

BEGIN WORK; --Explicit transaction started

UPDATE ...
UPDATE ... --AUTOCOMMIT OFF
COMMIT WORK;
--Explicit transaction committed
--AUTOCOMMIT STILL ON

INSERT ...; --Implicit transaction started

--Row committed
INSERT ...; --Implicit transaction started
. --Row committed
.
.

Cursor Operations (Embedded SQL)

If you are performing cursor UPDATE or DELETE operations and have the
AUTOCOMMIT option turned on, the transaction is committed and the cursor closed at
the end of the first UPDATE WHERE CURRENT OF or DELETE WHERE CURRENT
OF. You cannot perform another fetch without reopening the cursor. You should turn off
the AUTOCOMMIT feature for cursor operations. When you complete your cursor
operations, you can turn the AUTOCOMMIT feature on again.

CONTROL TABLE Directive

In NonStop SQL/MP, you can use the CONTROL TABLE directive for many purposes
but mainly to specify performance-related options for DML accesses to a table or view.
In NonStop SQL/MP, the CONTROL TABLE directive affects decisions the SQL
compiler makes about how to execute DML statements. In NonStop SQL/MX, because
the compiler and executor have been redesigned, many directives are no longer
necessary.
In NonStop SQL/MP, you can enter multiple options on a single CONTROL TABLE
statement. In NonStop SQL/MX, you can enter only one option on a single CONTROL
TABLE statement.
Table 3-5 lists which SQL/MP controls are not implemented in NonStop SQL/MX or are
implemented differently:

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 21
DML Features CONTROL TABLE Directive

Table 3-5. CONTROL TABLE Directives

Implemented
SQL/MP CONTROL in NonStop
TABLE Directives SQL/MX? Comments
ACCESS PATH Changed Access paths can be forced using CONTROL
QUERY SHAPE. See Forcing Query Execution
Plans on page 5-8.
JOIN METHOD Changed Join methods can be forced using CONTROL
QUERY SHAPE. See Forcing Query Execution
Plans on page 5-8.
JOIN SEQUENCE Changed Join sequence can be defined with a CONTROL
QUERY DEFAULT setting,
JOIN_ORDER_BY_USER. See CONTROL QUERY
Directive on page 3-23.
TIMEOUT Changed The SQL/MP CONTROL TABLE * TIMEOUT
statement combines the functions of the CONTROL
TABLE TIMEOUT and SET TABLE * TIMEOUT
statements.
OPEN PARTITIONS No
SEQUENTIAL No
SEQUENTIAL No
BLOCKSPLIT
SKIP No
UNAVAILABLE
PARTITION
SYNCDEPTH No

Use the MXCI command, SHOWCONTROL, to see which controls are in effect.
In NonStop SQL/MX, the CONTROL TABLE directive requires that you enclose all
arguments in single quotes. In NonStop SQL/MP, the CONTROL TABLE directive has
no such restriction.
In NonStop SQL/MP, the control TIMEOUT allows you to use a negative value (-1),
which is essentially wait forever. This value is not allowed in NonStop SQL/MX.
The NonStop SQL/MP CONTROL TABLE directive enables you to adjust the timeout
value for compile time. NonStop SQL/MX provides compile time directives for
CONTROL TABLE TIMEOUT and CONTROL TABLE STREAM_TIMEOUT. NonStop
SQL/MX provides for run-time timeout with the SET TABLE TIMEOUT statement as
described in the SQL/MX Reference Manual and SQL/MX Programming Guide for C
and COBOL.
See the SQL/MX Reference Manual for additional information about the CONTROL
TABLE directive.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 22
DML Features CONTROL QUERY Directive

CONTROL QUERY Directive

In NonStop SQL/MP, the CONTROL QUERY compiler directive controls plans for
queries. Options specify whether to resolve names at execution time or startup time,
whether to include hash join algorithms considered for executing queries, and whether
to optimize query response time for returning a few rows or all rows. In NonStop
SQL/MX, the CONTROL QUERY directive has been changed to CONTROL QUERY
DEFAULT. Many options have been changed.
The next table lists the SQL/MP CONTROL QUERY directives. More information about
default settings is provided in the SQL/MX Reference Manual and the SQL/MX Query
Guide.
How NonStop SQL/MX
SQL/MP CONTROL QUERY Affects the Same
Directive Functionality Comments
BIND NAMES Internally handled in the
compiler. No external
control.
HASH JOIN CONTROL QUERY SHAPE See Forcing Query
to force Execution Plans on
page 5-8.
INTERACTIVE ACCESS ON Not implemented.
MDAM CONTROL TABLE or ON by default; see Forcing
CONTROL QUERY SHAPE Query Execution Plans on
to force off page 5-8.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 23
DML Features Query Caching

Query Caching
In NonStop SQL/MP, you control statement and query caching with MXCS and
SQL/MP configuration attributes, system defaults, and control statements. These
settings persist until a user changes a Guardian user ID, a server is reinitialized, or a
rollback condition occurs. Caching takes place at a high level.
In NonStop SQL/MX, caching is controlled by control query default statements.
Caching takes place within each SQL/MX MXCMP compiler session and is applicable
to query compilation requests in that session. MXCS, JDBC, and MXCI clients have
their own MXCMP sessions. Each MXCMP session has caching turned on by default.
MXCS and JDBC/MX clients use control statements that you can store in a named
collection linked to a user and stored in configuration tables for that user, so they are
always set for that user.
See the SQL/MX Reference Manual for details about these settings and the SQL/MX
Query Guide for more information about queries.
This table compares caching features in NonStop SQL/MP and NonStop SQL/MX:

Feature NonStop SQL/MP NonStop SQL/MX

Cache location In ODBC/MP server In MXCMP compiler
Cache setting Disabled by default Enabled by default
Cache size limit 32,767 statements 4,194,303 KB of memory
Efficiency increase High Medium
from cache hit
Effect of schema Flushes cache. Flush and re-cache
change on cached transparently.
query
Caching Caching is discouraged for Complex queries and
considerations SELECT queries with DSS-type queries are
ORDER BY clauses. currently not cached.
An optional feature of
statement cache supports
caching parallel plans and
allocated ESPs.
To avoid resource
exhaustion, use caching only
where ESP sharing is
enabled and the number of
statements cached are small.
Location of caching ODBC/MP user's profile. SQL/MX’s MXCS session.
control

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 24
DML Features Query Caching

Feature NonStop SQL/MP NonStop SQL/MX

When caching On next system configuration At next query compile.
change takes effect refresh.
Ownership of object Cache per ODBC/MP server. Cache per MXCMP process.
When cache is When Guardian user ID On demand. User requests
flushed changes. CONTROL cache size be reduced to
statements only affect the zero by issuing CONTROL
next statement compilation, QUERY DEFAULT
which is avoided when a QUERY_CACHE '0'
statement is used from statement.
cache. Cache is not
preserved over critical
condition termination. All
statements are flushed at
server reinitialization.
Cache statistics Logged in MP table. In-memory counters.
When statistics are Performed by server, either In-memory counters are
saved and reported at statement close or at retrieved when user issues
session disconnect. SELECT … FROM
table(querycache())
WHERE …
and
SELECT … FROM
table(querycacheentri
es()) WHERE …
statements.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 25
DML Features Query Caching

Feature NonStop SQL/MP NonStop SQL/MX

Cache counters Number of statements, In-memory counters are:
lookups, hits, drops. Number of entries, pinned
Extended to indicate whether entries, compiles, cachable
statement was not cached compiles by phase,
because it is parallel, and recompiles, retries,
parallel was not activated. cachables, hits by phase,
pinned hits by phase,
displaced, cachable-but-too-
large, current size, max size,
average plan size, max
victims, optimization level,
pinning state.
Per entry: hits, size, text,
plan ID, row ID, phase,
optimization level, catalog
name, schema name, pin,
number of parameters,
parameter types, plan length,
compilation time, average hit
time, required shape,
isolation level, access mode,
autocommit, rollback mode
setting.
Types of queries that ExecDirect or Prepare user ExecDirect or Prepare user
are cachable SQL DML statements, SQL DML statements.
without dynamic parameters.
When constant Just before execute step in After generate step in
values are bound to ODBC/MP server. Literals MXCMP compiler.
a query's are transformed into
system-introduced parameters during the
parameters statement translate stage.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 26
DML Features Query Type Comparison

Query Type Comparison

Simple data manipulation language (insert, update, delete, select) statements without
predicates, or with key equality predicates that are guaranteed to affect at most one
row, are cached in both NonStop SQL/MP and NonStop SQL/MX. An equality
predicate is cachable if one of the operands is a literal and the other operand is a fully
specified clustering (or primary) key.
Data manipulation statements with inequality predicates, nonkey predicates,
subqueries, multirow and multiway joins are cached in NonStop SQL/MP but not in
NonStop SQL/MX. Many decision support type queries are cached in NonStop
SQL/MP but not in NonStop SQL/MX.
Table 3-6 describes various statements, with examples, and indicates whether they are
cached by NonStop SQL/MP and NonStop SQL/MX:

Table 3-6. Query Type Comparison (page 1 of 3)

NonStop NonStop
Type of Query SQL/MP SQL/MX
Tuple inserts: Y Y
INSERT INTO t1 VALUES(1,1,1)
Insert select with key equality predicate: Y Y
INSERT INTO t1 SELECT * FROM s
WHERE key_col_a=1
Insert select with column comparison predicate: Y N
INSERT INTO t1 SELECT * FROM t2
WHERE key_col_a=key_col_b
Insert select with any nonkey predicate: Y N
INSERT INTO t1 SELECT * from t2
WHERE nonkey_col<1
Insert select with subquery: Y N
INSERT INTO t1 SELECT * FROM t2
WHERE col_a IN(SELECT col_b FROM u)
Updates with key equality predicates: Y Y
UPDATE t1 SET col.a=1 WHERE key_col=2
Updates with column comparison predicate: Y N
UPDATE t1 SET col.a=1
WHERE key_col_a=key_col_b
Updates with any nonkey predicate: Y N
UPDATE t1 SET col.a=1 WHERE nonkey_col_a>2
Updates with subquery: Y N
UPDATE t1 SET a=1 WHERE col_a IN
(SELECT col_b FROM t2 WHERE col_c=1)

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 27
DML Features Query Type Comparison

Table 3-6. Query Type Comparison (page 2 of 3)

NonStop NonStop
Type of Query SQL/MP SQL/MX
Deletes with key equality predicates: Y Y
DELETE FROM t1 WHERE KEY=1
Deletes with column comparison predicate: Y N
DELETE FROM t1 WHERE key_col_a=key_col_b
Deletes with any nonkey predicate: Y N
DELETE FROM t1 WHERE nonkey_col_a <> 1
Deletes with subquery: Y N
DELETE FROM t1 WHERE col_a IN
(SELECT col_b FROM t2 WHERE col_c=1)
Selects with key equality predicate: Y Y
SELECT * FROM t1 WHERE key_col_a=1
Select with column comparison predicate: Y N
SELECT * FROM t1 WHERE key_col_a=key_col_b
Selects with a nonkey predicate: Y N
SELECT * FROM t1 WHERE nonkey_col<1
Selects with a subquery: Y N
SELECT * FROM t1
WHERE col_a IN (SELECT col_b FROM t2)
Single row two-way joins: Y Y
SELECT * FROM t1,t2
WHERE t1.key_col_a=t2.key_col_b
AND t1.key_col_a=1
AND t2.key_col_b=2
Selects with multirow two-way joins: Y N
SELECT * FROM t1,t2
WHERE t1.key_col_a=t2.key_col_b
Selects with multirow, multiway nonequi joins: Y N
SELECT * FROM t1,t2,t3
WHERE t1.key_col_a>t2.key_col_b
AND t2.key_col_b<t3.key_col_c
Aggregates with key equality predicate: Y Y
SELECT COUNT(*) FROM t1 WHERE key_col_a=1
Aggregates with column comparison predicate: Y N
SELECT COUNT(*) FROM t1
WHERE key_col_a=key_col_b
Aggregates with nonkey predicate: Y N
SELECT COUNT(*) FROM t1 WHERE nonkey_col<1

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 28
DML Features Versioning

Table 3-6. Query Type Comparison (page 3 of 3)

NonStop NonStop
Type of Query SQL/MP SQL/MX
Aggregates with subquery: Y N
SELECT COUNT(*) FROM t1
WHERE col_a IN (SELECT col_b FROM t2)
Most DSS queries: Y N
SELECT COUNT(*) FROM t1
WHERE col_c<=1 GROUP BY col_a
Statements that reference date-time data types Y N
Data Definition Language statements N N
Data Control Language statements N N

Versioning
NonStop SQL/MX database objects inherit their schema's version, regardless of what
features they use. SQL/MP database objects have a feature-based object version.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 29
DML Features Versioning

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

3- 30
 4 Embedded SQL
As in NonStop SQL/MP, to access an SQL/MX database, you can execute SQL
statements interactively by using MXCI (the NonStop SQL/MX equivalent to SQLCI) or
programmatically by embedding, or including, SQL statements in a host program. In
NonStop SQL/MX, you can write the host programs by using embedded SQL in the
ANSI C, C++, COBOL, or Java programming languages.
This section discusses differences in the use of embedded SQL in C, C++, and
COBOL programs between NonStop SQL/MP and NonStop SQL/MX. The main
differences are in the diagnostics areas for error handling and in the descriptor areas
for dynamic SQL programming. For detailed information about the use of embedded
SQL in C, C++, and COBOL programs in NonStop SQL/MX, see the SQL/MX
Programming Guide for C and COBOL. For information about embedding SQL/MX
statements in Java programs, see the SQL/MX Programming Manual for Java.
This section is organized into these categories:
Host Variable Declarations
Data Type Conversion
Host Variables With Date-time Data Types
Indicator Variables
Embedded SQL Statements
Dynamic SQL
Error Handling
Statistics Area
Program Development

Host Variable Declarations

NonStop SQL/MP allows the declaration of character host variables by using arithmetic
operations to determine the size of the host variable array. NonStop SQL/MX does not
allow arithmetic operations to determine the size of the array. In NonStop SQL/MX,
host variables must have constants specifying the size of the array.
Many SQL/MP programs declare host variables from structures that are defined
outside of the DECLARE SECTION. NonStop SQL/MX requires that structure
definitions for host variables must be declared between BEGIN DECLARE SECTION
and END DECLARE SECTION.

SETSCALE Function
In NonStop SQL/MP, using the C programming language, use the SETSCALE function
to set the scale of certain data types. You must use the SETSCALE function whenever
the host variable is used in the query. In NonStop SQL/MX, scale information for a host
variable is supplied when the host variable is declared inside the BEGIN DECLARE
SECTION.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

4 -1
Embedded SQL Variable-Length Character Data in Embedded SQL
C Programs

SQL/MP Example
This C program fragment uses SETSCALE with an INSERT statement to create a new
row with the value 98.34 in the PARTS.PRICE column after storing the value in host
variable unit_price. The value is multiplied by 100 for storing as a whole number.
unit_price = 9834;
EXEC SQL INSERT INTO =PARTS (PRICE)
VALUES (SETSCALE (:unit_price, 2));

NonStop SQL/MX Example

You can declare a host variable as NUMERIC (p,s), where p is the precision and s is
the scale. You do not use the SETSCALE function when you use the host variable. For
example, you can declare a numeric(4,2) host variable as:
NUMERIC(4,2) unit_price;
The same scale applies when a value is fetched. For this example, the fetched value
has an implicit scale of 2.

Variable-Length Character Data in Embedded SQL C Programs

In NonStop SQL/MP, the VARCHAR data type defined for a column represents one
data element. However, suppose the column CUSTNAME is defined as VARCHAR(26)
in NonStop SQL/MP. The SQL/MP INVOKE directive generates a structure of the form:
struct
{
short len;
char val[27];
} custname; /* SQL/MP variable-length character data */
where len is a numeric data item that represents the length, and val is a fixed-length
character data item for the string plus an extra byte for the null terminator. In NonStop
SQL/MP, you can also explicitly declare a structure of the preceding form as a host
variable for a VARCHAR column.
In NonStop SQL/MX, in addition to the char data type, you can declare the VARCHAR
data type for host variables within embedded SQL C programs (SQL:1999 provides the
VARCHAR data type for C programs). For example, this declaration is for a column up
to 26 bytes:
VARCHAR custname[27];/* SQL/MX variable-length character data */
As in NonStop SQL/MP, the last byte of the array is for the null terminator. You must
declare the array one byte larger to allow for the null terminator. The null terminator
follows the data inserted, so if you insert a six-character string into a VARCHAR host
variable specified as 27 characters, the null terminator is the seventh character.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

4 -2
Embedded SQL INVOKE Directive

INVOKE Directive
In NonStop SQL/MP, use the INVOKE directive interactively through SQLCI to create
host variable declarations in an EDIT file. The EDIT file can then be included in an
embedded SQL program.
In NonStop SQL/MX, use the INVOKE directive to create host variable declarations
only directly in an embedded SQL program. Use the INVOKE command in MXCI to
display the table and column names and data types associated with the columns; you
cannot invoke the format of another language (such as COBOL).
For either kind of table, INVOKE does not support the DATEFORMAT clause. INVOKE
returns the ANSI DATETIME data type. The SQL DATEFORMAT function returns a
character string. They are not compatible and cannot be used interchageably.

Data Type Conversion

For information about the valid conversions you can perform with CAST, see the
SQL/MX Reference Manual.

Host Variables With Date-time Data Types

In NonStop SQL/MX, you can convert data to the data type you specify by using the
CAST specification for any expression, whether or not you are using a parameter.
NonStop SQL/MX replaces the SQL/MP TYPE AS clause with the CAST specification.
In previous releases of NonStop SQL/MX, you were required to cast DATE, TIME,
TIMESTAMP, and INTERVAL data types. In NonStop SQL/MX Release 2.0 you can
directly declare them as host variables.
In NonStop SQL/MP, use the CAST function to associate a character or numeric data
type with a dynamic input parameter. Use the TYPE AS clause in the parameter
specification to associate a date-time or interval data type with a parameter.

Indicator Variables
As in NonStop SQL/MP, the primary use of the indicator variable in NonStop SQL/MX
is to indicate whether the value supplied is null. If the source value is null, the value of
the indicator variable is set to less than zero.
However, in NonStop SQL/MX, the indicator variable has another use. If an indicator
variable is associated with a target character host variable in the INTO clause of a
SELECT or FETCH statement, and the length of the target variable is smaller than the
value returned, the truncated string is placed into the host variable, the indicator
variable is set to the length of the source string, and a warning message is issued.
Otherwise, if the length of the target variable is sufficient, the indicator variable is set to
0.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

4 -3
Embedded SQL Embedded SQL Statements

For example, suppose the columns PARTDESC and QTY_AVAILABLE in the PARTS
table allow null. After the SELECT statement executes, this C example tests the
indicator variable for null or a truncated value:
EXEC SQL SELECT partnum, partdesc, price, qty_available INTO
:parts_rec.partnum,
:parts_rec.partdesc
INDICATOR :parts_rec.partdesc_i,
:parts_rec.price,
:parts_rec.qty_available
INDICATOR :parts_rec.qty_available_i,
FROM sales.parts
WHERE partnum = 300380 AND partnum = 2402 ;
...
if (parts.qty_available_i < 0) handle_null_value();
if (parts.partdesc_i > 0 ) handle_truncated_value();
...
In NonStop SQL/MP, the indicator variable is not set to the source length in the case of
truncation. It is set to 0 if it is a nonnull value and if the value is null, the indicator
variable is set to -1.

Embedded SQL Statements

Most SQL/MP embedded SQL statements exist in NonStop SQL/MX. However, a few
of the statements, options, and clauses have been dropped because of SQL:1999
ANSI changes.

Table 4-1. Embedded SQL Statements

Implemented
SQL/MP Embedded SQL in NonStop
Statements and Directives SQL/MX? Comments
CLOSE Yes
DECLARE CURSOR Changed See Updatable Cursors.
DESCRIBE Changed NonStop SQL/MX uses an SQL descriptor
area, as defined by ANSI SQL:1999. See
Descriptor Area.
DESCRIBE INPUT Changed NonStop SQL/MX uses an SQL descriptor
area, as defined by ANSI SQL:1999. See
Descriptor Area.
EXECUTE Changed No RETURNING clause in NonStop
SQL/MX.
EXECUTE IMMEDIATE Yes
FETCH Changed NonStop SQL/MX requires the number of
host variables in the fetch list to be the same
as the number of values returned in the
cursor specification.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

4 -4
Embedded SQL EXECUTE Statement

Table 4-1. Embedded SQL Statements

Implemented
SQL/MP Embedded SQL in NonStop
Statements and Directives SQL/MX? Comments
FREE RESOURCES No
INCLUDE STRUCTURES No See Descriptor Area, Diagnostics Area,
Statistics Area.
INCLUDE SQLCA No See Diagnostics Area.
INCLUDE SQLDA No See Descriptor Area.
INCLUDE SQLSA No See Statistics Area.
INSERT Changed No RETURNING clause in NonStop
SQL/MX. See INSERT Statement on
page 3-14.
INVOKE Changed No FORMAT clause in NonStop SQL/MX.
See INVOKE Directive.
OPEN Changed NonStop SQL/MX uses an SQL descriptor
area, as defined by ANSI SQL:1999. See
Descriptor Area.
Positioned DELETE Yes
Positioned UPDATE Changed NonStop SQL/MX allows this statement to be
dynamic.
PREPARE Yes SQL/MX FROM clause does not allow
quoted strings.
RELEASE No
SELECT INTO Changed See SELECT Statement.
WHENEVER Changed Usage changes. See WHENEVER
Declarative.
Positioned DELETE Yes

EXECUTE Statement
In NonStop SQL/MX, input data is supplied by the USING clause that specifies either a
list of host variables or an SQL descriptor area. Output data is provided by the INTO
clause that specifies either a list of host variables or an SQL descriptor area. See
Descriptor Area.
NonStop SQL/MP implements the RETURNING clause for output data. NonStop
SQL/MX does not provide a RETURNING clause.
In both NonStop SQL/MX and NonStop SQL/MP, if you perform a PREPARE and an
EXECUTE of an SQL UPDATE, INSERT or DELETE statement on a non-audited table,
NonStop SQL/MX does not close the table and will not release record and table locks.
You can use FUP LISTLOCKS to check if locks are kept on that table by your MXCI
session after completion of the EXECUTE statement, then manually execute an

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

4 -5
Embedded SQL Updatable Cursors

UNLOCK command or exit the MXCI session. You can also use standalone UPDATE,
INSERT or DELETE statements instead of using PREPARE and EXECUTE. Only
SQL/MP tables can be non-audited.

Updatable Cursors
In NonStop SQL/MP, delete operations using a cursor do not require the FOR UPDATE
OF clause of the DECLARE CURSOR statement; you must use this clause when you
update rows, but it is optional when you delete rows. In NonStop SQL/MX, both update
and delete operations require an updatable cursor.
In NonStop SQL/MP, the cursor defaults to read only unless specified FOR UPDATE.
In NonStop SQL/MX, within the DECLARE CURSOR statement, the optional FOR
clause has this form:
FOR {READ ONLY | UPDATE [OF colname [,colname]...]}
By using the FOR clause, you can specify whether the cursor is FOR READ ONLY
(read-only cursor) or FOR UPDATE OF (updatable cursor). If you do not specify the
FOR clause, and if ORDER BY or GROUP BY is specified for the cursor or if the query
expression defining the cursor specifies a read-only table, the cursor defaults to
read-only. It is the query that determines that the table is read-only, not the table’s
attributes. Otherwise, the cursor defaults to updatable (without a column list). If you
know that you are not going to update a cursor, specify FOR READ ONLY.

FETCH Statement
In NonStop SQL/MP, when a FETCH statement executes, if the number of host
variables is different from the number of columns in the result table of the declared
cursor, a warning is issued, and NonStop SQL/MP returns the number of values in the
shorter list.
In NonStop SQL/MX, the number of host variables must be the same as the number of
columns in the result table or an error is returned.

Positioned UPDATE Statement

In NonStop SQL/MP, you can execute the UPDATE ... WHERE CURRENT OF
statement for static SQL only.
In NonStop SQL/MX, you can execute the UPDATE ... WHERE CURRENT OF
statement by static SQL and dynamic SQL.

Dynamic SQL
Dynamic SQL allows a program to construct, compile, and execute all or part of an
SQL statement at run time. NonStop SQL/MP provides a descriptor area, referred to as
the SQLDA, containing information about input parameters and output variables in
dynamic SQL statements.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

4 -6
Embedded SQL Descriptor Area

NonStop SQL/MX provides for the allocation of a descriptor area with the use of
SQL:1999 dynamic SQL statements. This descriptor area consists of multiple item
descriptor areas, together with a count of the number of those item descriptor areas.
For further information about using the descriptor area, see Descriptor Area on
page 4-7.
The next list summarizes the dynamic SQL statements that you use with an SQL/MX
descriptor area. For information about the syntax and use of these statements, see the
SQL/MX Reference Manual.

ALLOCATE DESCRIPTOR Allocates an input or output SQL descriptor area.

DEALLOCATE DESCRIPTOR Deallocates an SQL descriptor area.
GET DESCRIPTOR Retrieves information from an SQL descriptor
area—the COUNT of the item descriptor areas and
specified fields in the areas.
SET DESCRIPTOR Modifies information in an SQL descriptor area.
PREPARE Prepares (compiles) a dynamic SQL statement for
subsequent execution by an EXECUTE statement.
DEALLOCATE PREPARE Deallocates a prepared statement and returns the
system resources used by the statement; also
allows you to reuse the name of the statement.
DESCRIBE [OUTPUT] Stores in the output descriptor area information
about output values (usually SELECT columns)
from a prepared statement.
DESCRIBE INPUT Stores in the input descriptor area information
about dynamic input parameters for a prepared
statement.
EXECUTE Executes a prepared dynamic SQL statement.
EXECUTE IMMEDIATE Prepares (compiles) and executes a dynamic SQL
statement.

Descriptor Area
NonStop SQL/MP provides a descriptor area (SQLDA) that enables a dynamic SQL
application to exchange information with the database about input parameters and
output columns. You can declare the descriptor area by using the INCLUDE SQLDA
directive.
NonStop SQL/MX does not provide the SQLDA as does NonStop SQL/MP and does
not provide the INCLUDE SQLDA directive. However, NonStop SQL/MX does provide
an SQL descriptor area for dynamic SQL, which consists of multiple item descriptor
areas, together with a count of the number of those item descriptor areas.
In NonStop SQL/MX, you must allocate and deallocate input and output SQL
descriptor areas by using the ALLOCATE DESCRIPTOR and DEALLOCATE

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

4 -7
Embedded SQL Error Handling

DESCRIPTOR statements. After the parameters have been described (by using the
DESCRIBE statement as in NonStop SQL/MP), you can modify and retrieve
information from the descriptor areas only by using the SET DESCRIPTOR and GET
DESCRIPTOR statements, respectively.
If you have dynamic input parameters in a prepared SQL statement, you must code the
appropriate 3GL statements and use SET DESCRIPTOR to set the values in the
descriptor area. If you have more than one dynamic input parameter, you can identify
the values by their position in the prepared SQL statement. If you are using descriptor
areas for input parameters, you must use the CAST specification to specify the
appropriate data type.
In NonStop SQL/MX, you should give input parameters a defined data type provided in
the descriptor area by using the CAST specification. The CAST specification is not
needed for MXCI parameters or for parameters used in dynamic cursors in embedded
SQL programs.
For examples that use dynamic SQL statements, see the SQL/MX Reference Manual.

Error Handling
The major difference in error handling between NonStop SQL/MP and NonStop
SQL/MX relates to the SQL:1999 standard, as described next.

Format of Error Codes

NonStop SQL/MP provides an integer-valued SQLCODE as the mechanism to send
error information from the database to the application.
NonStop SQL/MX supports both an SQLCODE mechanism and the new SQLSTATE
codes, as specified in the SQL:1999 standard. However, the SQLCODE values
generated by NonStop SQL/MX are different from those generated by NonStop
SQL/MP. For portable applications, you should use SQLSTATE. Because of the
overhead associated with SQLSTATE, you will most likely use it only for error handling
in ANSI-compliant applications. For information about SQLSTATE and error handling in
NonStop SQL/MX, see the SQL/MX Programming Guide for C and COBOL.

Equivalent SQLCODE and SQLSTATE Values

NonStop NonStop NonStop

SQL/MP SQL/MX SQL/MX
Error SQLCODE SQLCODE SQLSTATE
SQL_NO_ERROR 0 0 00000
SQL_NO_ROWS_MATCH 100 100 02000
SQL_TOO_MANY_MEET_CRITERIA -8222 -8401 21000
SQL_DUPLICATE_ENTRY -8227 -8102 23000

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

4 -8
Embedded SQL WHENEVER Declarative

SQL_PARTITION_NOT_AVAIL -8239 * *
SQL_FILE_SYSTEM_ERROR -8300 -8551 -
SQL_NO_INDICATOR_VAR -8423 -8420 22002
SQL_LOST_OPEN_WARN 8204 8574 -
SQL_LOST_OPEN_ERROR -8204 -8574 -

* Not available. The current release of NonStop SQL/MX does not support the SQL/MP
CONTROL TABLE SKIP UNAVAILABLE PARTITION option.

WHENEVER Declarative
In NonStop SQL/MP, the WHENEVER declarative specifies an action to take when an
error, warning, or no-rows-found condition occurs. Further, NonStop SQL/MP supports
all three actions: CONTINUE, GO TO (or GOTO), and PERFORM (in COBOL) or
CALL (in C).
NonStop SQL/MX continues to support all three conditions, although SQL:1999
supports only SQLERROR and NOT FOUND.
In addition, NonStop SQL/MX continues to support all three actions, although
SQL:1999 supports only CONTINUE and GO TO. Without PERFORM and CALL, the
WHENEVER declarative is inadequate; the problem with supporting only CONTINUE
and GO TO is that a program must pass control to a named procedure. The
assumption is that an error is reported by the procedure, and the application halts
execution. However, in many situations, an application can recover from errors.
Therefore, your application should execute an error-handling procedure and pass
control back to the main program.
For information about handling errors, see Diagnostics Area, next.
In NonStop SQL/MP, the WHENEVER declarative requires a colon (:) before the target
location for the GO TO, PERFORM (in COBOL) or CALL (in C) actions. This restriction
is removed in NonStop SQL/MX.

Diagnostics Area
NonStop SQL/MP provides a communication area (SQLCA) that enables an
application program to check the status of SQL statement execution. You can declare
the communication area by using the INCLUDE SQLCA directive.
NonStop SQL/MX does not provide the SQL communication area or use the INCLUDE
SQLCA directive. Instead, NonStop SQL/MX follows the ANSI SQL:1999 standard by
automatically providing a diagnostics area that stores completion and exception
information.
At the beginning of the execution of an SQL statement, the diagnostics area is emptied
by the preprocessor. When the statement executes, NonStop SQL/MX places
completion and exception information in this area. You can access the information in
the diagnostics area by using the GET DIAGNOSTICS statement.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

4 -9
Embedded SQL Statistics Area

Suppose your application has executed an SQL statement and the WHENEVER
statement reports an error. To find out more about the results of the executed
statements, code your application to execute a procedure that uses GET
DIAGNOSTICS:
For example, in a C program, the code might look like this:
/* First, get the count of the number of exception conditions.
*/
EXEC SQL GET DIAGNOSTICS :hv_num = NUMBER,
:hv_cmdfcn = COMMAND_FUNCTION,
:hv_dynfcn = DYNAMIC_FUNCTION,
...;
/* Second, get the i-th condition and process it. */
for (i = 1; i <= hv_num; i++) {
EXEC SQL GET DIAGNOSTICS EXCEPTION :i
:hv_tabname = TABLE_NAME,
:hv_colname = COLUMN_NAME,
:hv_sqlstate = RETURNED_SQLSTATE
...;
... /* Process the SQL error information. */
}

Statistics Area
NonStop SQL/MP provides a performance and statistics area (SQLSA) that enables an
application to access this type of information about DELETE, INSERT, SELECT,
UPDATE, OPEN, FETCH, DESCRIBE, and PREPARE statements. NonStop SQL/MX
currently has no structure that corresponds to SQLSA and cannot provide information
on what type of statement was compiled or on the cost of a PREPARE statement.

Program Development
SQL/MP Compiler
Compiling programs using the SQL/MP compiler differs from using the SQL/MX
compiler. In NonStop SQL/MP, compiling a host language program consists of these
steps:
1. Create the host language source file with embedded SQL statements and
directives.
2. Run the C or COBOL compiler to compile the statements in the source file.
3. Run the Binder program if necessary to bind object files.
4. Run the SQL compiler (SQLCOMP).
The result is a valid SQL program file, registered in the catalog and ready for
execution. Graphically, this procedure appears similar to Figure 4-1.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

4- 10
Embedded SQL SQL/MX Compiler

Figure 4-1. SQL/MP Compilation Process

Host Language and

Source File SQL Statements

Binder Process to
Combine the Host
Language Object File
Host Language
With Other Object Files
Compiler

Binder Process

Host
Language
Object File
Host Language Object
File With SQL Source
Statements
SQL
Compiler

SQL
Valid SQL Program File
Program File
Ready for Execution

SQL/MX Compiler
The SQL/MX program compilation process is different from that of NonStop SQL/MP.
After you have compiled your program, you have two files: a host language source file
and an SQL/MX module.
This high-level procedure compiles a program with NonStop SQL/MX and is shown in
Figure 4-2 on page 4-12:
1. Using an editor, create the embedded SQL application.
2. Run the SQL/MX host language (C or COBOL) preprocessor to:
• Parse the EXEC SQL statements and replace them with CLI calls.
• Create a module definition file describing the SQL statements. The module
definition file is a separate file instead of being embedded within the compiled
objects.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

4- 11
Embedded SQL SQL/MX Compiler

3. Run the SQL/MX compiler to create execution plans for the SQL statements and
store the plans in a module file.
4. Run the C or COBOL compiler to compile the statements in the source file.
5. Execute the program in the OSS environment, as described in the SQL/MX
Programming Guide for C and COBOL.

Figure 4-2. SQL/MX Compilation Process

Source File Host Language and

(text file) SQL Statements

NonStop SQL/MX
Host Language
Preprocessor

SQL Module
Host Language Host Language
Source File Definition File
Statements (temporary text file) SQL Statements
(temporary text file)

Host Language NonStop SQL/MX

Compiler Compiler

Host Language SQL Statements

Host Language SQL Module
Program
Object Code (stored in tdm_sqlmodules)
and Object Code

The key difference in the SQL/MX compilation is the resulting two files: the SQL
module, which stores the query plan, and the host language program file.
For more information about the SQL/MX module, see the SQL/MX Programming Guide
for C and COBOL. For syntax for the MODULE directive, see the SQL/MX Reference
Manual.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

4- 12
Embedded SQL Stored Procedures in Java

Stored Procedures in Java

Both NonStop SQL/MP and NonStop SQL/MX allow you to use stored procedures, with
these differences:

SQL/MP Stored Procedures

• Only callable from ODBC/MP.
• The called stored procedure must be a Pathway server with embedded SQL,
written in COBOL, C, or TAL.
• Cannot use SQL/MX features like Publish/Subscribe.

SQL/MX Stored Procedures

• Callable from any SQL/MX application: MXCS, JDBC/MX, NonStop SQLJ, MXCI,
or embedded SQL(MX) within a COBOL or C/C++ program.
• The called stored procedure must be written in Java.
• Implementation closely follows ANSI standards (SQL99 and SQL/JRT).
For more details on SQL/MX stored procedures, see the SQL/MX Guide to Stored
Procedures in Java.

Similarity Checks and Automatic

Recompilation
NonStop SQL/MX and SQL/MP perform similarity checking in different ways. These
are differences between the products:
Difference between compile and run Does SQL/MP Does SQL/MX
time module similarity check pass? similarity check pass?
Name of table Yes Yes
Data in the table Yes Yes
Number of partitions Yes for serial plans No, unless plan
No for parallel plans chooses online
transaction optimization
Partition key ranges Yes for serial plans Yes
No for parallel plans
Non-partitioned table at compile time Yes No, unless plan
is partitioned at run time chooses online
transaction optimization
ESP Parallelism used in the plan No No

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

4- 13
Embedded SQL Similarity Checks and Automatic Recompilation

Difference between compile and run Does SQL/MP Does SQL/MX

time module similarity check pass? similarity check pass?
Number of indexes - more indexes at Yes No for INSERT,
run time than at compile time UPDATE, DELETE
queries.
Yes for SELECT.
Number of indexes - fewer indexes at Yes, if index is not used No for INSERT,
run time than at compile time by plan UPDATE, delete
queries.
Yes, for SELECT, if
index is not used by
plan.
Unique index present both at compile Yes No, fails for SELECT
and run time
Key tags (or values) for indexes Yes No
Creating and redefinition timestamp Yes Yes
Audit attribute Yes, with minor No
exception
Any of these attributes: ALLOCATE, Yes Yes
AUDITCOMPRESS, BUFFERED,
CLEARONPURGE, EXTENT (primary
and secondary), LOCKLENGTH,
MAXEXTENTS, NOPURGEUNTIL,
OWNER, SECURE, SERIALWRITES,
TABLECODE, VERIFIEDWRITES
Statistics on the table Yes Yes
Column headings Yes Yes
Catalog where table is registered Yes Yes, could be in
different schema
Number of columns - more columns at Yes, with same No
run time that at compile time attributes and no
SELECT
VIEWs on the table Yes, only for qualifying No
protection views
Check constraints, referential integrity Yes. Ignores check No
constraints, or triggers on the table constraints

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

4- 14
 5 Optimizer and Executor
Both the compiler and the executor for NonStop SQL/MX are different from NonStop
SQL/MP. Some highlights of the new compiler include:
• Rule and cost-based optimizer
• Optimization of complex relational expressions
• Use of histogram statistics
• Extensible cost model
Some highlights of the rewritten executor include:
• Data-flow, scheduler-driven task model
• Increased parallelism

Optimizer
The optimizer is the stage in the SQL compiler that optimizes the query and
determines the optimal query execution plan. Many design differences in NonStop
SQL/MX occur before optimization. During the normalization phase, these tasks are
performed:
• Subquery transformation into joins
• Constant folding (for example, b < 0 * c becomes b < 0)
• Predicates pushed down or pulled up
• Other unconditional transformations (for example, left joins transformed to inner
joins)
This table highlights the differences between the SQL/MP and SQL/MX optimizers:

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

5 -1
Optimizer and Executor Optimizer

SQL/MP Optimizer SQL/MX Optimizer

Search engine uses dynamic programming Search engine uses branch and bound
(bottom-up optimization) to enumerate programming (top-down optimization).
possible join orders. Considers Cartesian Multipass design; in the pilot phase, tables are
products as part of the join enumeration. ordered in ascending order, and large tables
are scanned last.
Optimizations considered include: Optimizations considered include:
• Three-join algorithms (nested join, • Four-join algorithms (nested join, ordered
hybrid hash join, and merge join). hash join, hybrid hash join, and merge
• Parallel join algorithm (matching, join). Within a join, also supports range
repartitioning.
nonmatching).
• Repartitioned join. • Extensible query tree allows for parallel
operations anywhere in the query tree.
• Aggregation and grouping in DAM
• Repartitioned join.
(aggregate and sort GROUP BY).
• Unions cannot be parallelized. • Aggregation and grouping can be
executed in DAM and ESPs.
• Transitive closure; equality predicates
• Unions can be parallelized.
are rewritten to include implied
predicates. • Transitive closure more extensive;
extends to grouping columns, order by
columns, and column references in the
select list.
• GROUP BY operations can occur almost
anywhere in the query tree.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

5 -2
Optimizer and Executor Executor

SQL/MP Optimizer SQL/MX Optimizer

Costing and statistics: Costing and statistics:
• Uses captured statistics to estimate • Histograms are used to represent data
the number of rows. Consists of a distribution, not just second high and
number of values—UEC, second high second low values. Data distribution is
and second low values—on a per- divided into intervals. For each interval,
column basis. NonStop SQL/MX stores the number of
• Numeric data requires data be values and the UEC.
uniformly distributed (between the • Multicolumn UEC and row count.
second high and second low values).
• Histograms can capture the true
• Costing based on simple arithmetic. distribution, given enough histogram
• Cost is represented by a single intervals.
floating point value. • Costing aligned to concurrency and
parallelism.
• Cost is represented as cost vectors
composed of resource units.
Influencing query plan: Influencing query plan:
• Force plans through the CONTROL • Force plans through the CONTROL
directive (table scope). QUERY SHAPE statement (query tree
• Change a few defaults with CONTROL scope).
directives on a table basis. • Change user defaults in several ways: per
• Update statistics. query tree, table basis, system or session
basis.
• Modify the table, such as adding
• Update statistics.
indexes.
Transformations within queries: Optimizer internally handles transformations
• User must know how the optimizer within queries. Some transformations include:
handles predicates so that proper • Unconditional predicate transformations
query design yields an optimal plan. (for example, constant folding).
• Some predicate transformations done. • Subqueries transformed to joins.
• Predicates are pushed down as far as • Predicates rewritten and pushed down as
possible. far as possible (more than in NonStop
SQL/MP).
• Syntactic and semantic sort elimination.

Executor
The executor in NonStop SQL/MX has also been rewritten and takes advantage of
these features:
• Completely no-waited operations
• More work done in the DAM
• Multiple forms of parallelism
• Scalability through parallelism on multiple levels

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

5 -3
Optimizer and Executor SQL/MP Iterator Model

This table highlights the differences between the SQL/MP executor and the SQL/MX
executor.

SQL/MP Executor SQL/MX Executor

Control flow iterator model. See SQL/MP Data-flow, scheduler-driven task model. See
Iterator Model. SQL/MX Task Model.
Nested join, hash join, and merge join Nested join, hash join, and merge join
supported. supported. Within a join, also supports range
partitioning.
Aggregate and sort GROUP BY queries Aggregate, sort and hash GROUP BY queries
can be executed in DAM. can be executed in DAM.
Parallel join, GROUP BY, sort operations Increased parallelism supported through no-
supported through ESPs. wait interface to DAM (no ESPs needed for
some parallel plans).
Result data from DAM and inserted rows Request and result data to and from DAM and
can be buffered in larger messages inserted rows are always buffered.
(VSBB, RSBB, and Insert VSBB).
Uses ESPs to keep multiple DAM Does not require 1:1 mapping between ESPs
processes active at the same time. and DAMs of a partitioned table.
Requires a 1:1 mapping between ESPs
and DAMs of a partitioned table.

SQL/MP Iterator Model

The SQL/MP executor uses an iterator model, also called a control flow model. In the
iterator model, a query plan is executed by a series of function calls. Each function
implements an operator in the query plan. Because the implementation is through a
series of function calls, no pipelining or parallel processing is possible.
The interface between the SQL/MP executor and DAM is synchronous. The executor
process is blocked while it waits for a reply from a DAM. NonStop SQL/MP uses ESPs
to keep multiple DAMs active at the same time. NonStop SQL/MP requires a 1:1
mapping between ESPs and the DAMs of a partitioned table.
The next figure shows an example of the iterator model. This iterator model is a
serialized type of model, meaning that only one function is active at a time. The
numbers in the figure relate to these operations:
1. The join calls the scan function.
2. The scan asks DAM to retrieve rows.
3. Rows are returned from DAM to the scan.
4. Rows are returned from the scan function call to the join.
5. The join calls the scan function, providing as input the row produced by Step 4.
6. The scan asks DAM to retrieve all rows matching the input.
7. Rows are returned from DAM to the scan.
8. Rows are returned from the scan function to the join.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

5 -4
Optimizer and Executor SQL/MX Task Model

Nested Join

1 8
4 5

Scan A Scan B
2 7 Function call
Function call
for scan to 3 6 for scan to
DAM DAM

SQL/MX Task Model

NonStop SQL/MX uses a data flow and scheduler-driven task model to execute
queries. After a query is optimized, the optimizer generates an optimized, executable
query plan that goes to the executor. The executor creates an operator tree based on
the executable query plan and prepares a node for each operator.
The next figure shows an operator tree in the task model. Each operator is an
independent task, and data flows between operators through in-memory queues or by
interprocess communication. The queue pair operates between two operators. Queues
between tasks allow operators to exchange multiple requests or result rows at a time.
A scheduler coordinates the execution of tasks and runs whenever it has data on one
of its input queues.

Join Tasks

Queues

Scan Scan

Building on the previous figure, the next figure shows the steps that the executor might
take to process this simplified query tree. The numbers in the figure relate to the
operations:
1. The join places a request in the down queue of the left scan.
2. The scan places a group of rows in its up queue, which the join retrieves.
3. The join places a request in the down queue of the right scan. At the same time,
the join can place a request in the down queue of the left scan.
4. The right scan places a group of rows in its up queue, which the join retrieves. If
active, the left scan also places a group of rows in its up queue, which the join
retrieves.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

5 -5
Optimizer and Executor SQL/MX Task Model

The initial request from the join to the left scan queue must occur before the other
operations can begin. After the initial request is processed, both the left and right scans
can be processing information through their queues simultaneously.

Join

1 2 3 4

Scan Scan

The task model makes it easy to perform all internal operations asynchronously so that
a single server thread can have multiple I/Os outstanding. This model also allows
parallelism for both shared memory and distributed memory architectures. In-memory
queues are used for communication, while exchange operators are used for distributed
memory.

Parallelism
In NonStop SQL/MP, parallelism works only on a table that is already partitioned and
only for queries that retrieve data from multiple partitions. NonStop SQL/MX considers
partitioned parallelism in all cases. In addition, NonStop SQL/MX uses pipelined and
independent parallelism.
NonStop SQL/MP does not recognize multilevel parallelism. In NonStop SQL/MP,
either the entire query tree runs in parallel or none of it runs in parallel. NonStop
SQL/MX supports multilevel parallelism, which allows each operator to have a different
level or degree of parallelism, including none.
NonStop SQL/MP uses only ESP parallelism and needs one ESP per base table
partition. The number of ESPs in NonStop SQL/MX is determined more by the number
of available CPUs than it is by the number of partitions in the tables. If a table has
more partitions than available CPUs, the optimizer can group the partitions so that
each ESP processes multiple partitions as if they were a single partition. Each ESP will
group adjacent partitions.
NonStop SQL/MX can merge sorted parallel streams with some restrictions. For
example, a parallel merge join with an ORDER BY clause might not require a final sort.
In SQL/MP, multiple streams must be sorted.
For a full explanation of parallelism in NonStop SQL/MX, see the SQL/MX Query
Guide.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

5 -6
Optimizer and Executor Ways to Influence the Query Execution Plan

Ways to Influence the Query Execution Plan

In NonStop SQL/MP, you can write queries in different ways to affect how a query is
optimized. You must know how the SQL/MP optimizer handles predicates so that a
proper query design yields an optimal plan. You can write queries in such a way as to
eliminate certain types of joins. Query performance in NonStop SQL/MX is not affected
by query design, because the SQL compiler performs the majority of transformations
internally.
This list indicates a few actions you can take to influence a query execution plan in
NonStop SQL/MX:
• Setting default values. See Default Settings.
• Updating statistics regularly.
• Forcing query execution plans with the CONTROL QUERY SHAPE statement.
For detailed information about improving query performance, setting default values,
updating statistics, and forcing query execution plans, see the SQL/MX Query Guide.

Default Settings
In NonStop SQL/MP, you specify certain default settings through the CONTROL
directives. For example, the CONTROL EXECUTOR directive allows you to specify
whether you want parallel processing on or off. With the CONTROL TABLE directive,
you can specify that you want to use a certain join method or change the timeout
setting. With the CONTROL QUERY directive, you can adjust other default values.
In NonStop SQL/MX, the SYSTEM_DEFAULTS table is a user table that enables you
to specify default settings for options and other attributes when executing MXCI
commands and SQL queries that can be run through MXCI or within an application.
For some attributes, you can override a default value by specifying an option with the
statement or command. You might find it more efficient, however, to configure the
SQL/MX execution environment with default values. NonStop SQL/MX lets you change
the default settings of many more items than NonStop SQL/MP.
If you do not enter an option when entering a command or if you do not provide some
attribute associated with the execution of queries, NonStop SQL/MX chooses a default.
If you have specified system default settings for the execution environment, these
value are used. Otherwise, the SQL/MX default values are used.
In addition to configuring the execution environment with default values through the
defaults tables, in NonStop SQL/MX you can use the CONTROL TABLE and
CONTROL QUERY DEFAULT settings to change default values at run time. For more
information about these commands, see Section 3, DML Features and the SQL/MX
Reference Manual.
NonStop SQL/MX stores system defaults information in the SQL/MP user table named
DEFAULTS, which is registered in the system catalog. For a full listing of default values
that can be changed, see the SYSTEM_DEFAULTS Table entry in the SQL/MX
Reference Manual. See also the SQL/MX Query Guide.
HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003
5 -7
Optimizer and Executor EXPLAIN Features

This table lists some of the differences in default values between NonStop SQL/MP
and NonStop SQL/MX. If you want to mimic an SQL/MP system using NonStop
SQL/MX, you need to reset some default values to get the NonStop SQL/MP behavior.

Default NonStop SQL/MP NonStop SQL/MX

ISOLATION_LEVEL STABLE ACCESS READ_COMMITTED
PARALLEL_EXECUTION OFF ON
READ_ONLY CURSOR ON OFF
RECOMPILATION_WARNINGS ON OFF
MDAM OFF ON

EXPLAIN Features
Both NonStop SQL/MP and NonStop SQL/MX utilize EXPLAIN features to present
information about query execution plans. In SQL/MP, EXPLAIN is a directive or SQL
utility that describes the execution plan for queries.
In NonStop SQL/MX, you review query execution plans and query performance by
using the EXPLAIN function, DISPLAY_EXPLAIN command, and Visual Query Planner
application. The EXPLAIN function and DISPLAY_EXPLAIN command present query
execution plans in machine-readable format. The Visual Query Planner presents query
execution plans in a graphical user interface through MXCS. Detailed information
about each method is presented in the SQL/MX Query Guide.

Forcing Query Execution Plans

In NonStop SQL/MP, you force query execution plans with options on the CONTROL
TABLE directive on a per table basis. In NonStop SQL/MX, you force query execution
plans by using the CONTROL QUERY SHAPE statement to force the entire plan, not
just a particular table.
The EXPLAIN function and the CONTROL QUERY SHAPE statement use similar
identifiers for the nodes of an operator tree. You can also generate and force query
execution plans through the Visual Query Planner.
Usually, the SQL/MX optimizer presents the best query execution plan. On rare
occasions, you might want to revise the plan presented by the optimizer. Forcing the
plan means that you take control for your query execution plan away from the
optimizer.
To force a plan, you must first run the EXPLAIN function to generate the result table of
the EXPLAIN function for your DML statement. Next, review the result to see if the plan
you received was the plan you wanted. If you determine that you want to change the
plan, use the CONTROL QUERY SHAPE statement to force the plan.
For additional information about forcing query execution plans, see the SQL/MX Query
Guide.
HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003
5 -8
Optimizer and Executor Statistics

Statistics
In NonStop SQL/MP, the UPDATE STATISTICS statement updates the statistics stored
in the catalog for the specified table. In NonStop SQL/MX, the UPDATE STATISTICS
statement updates the histogram statistics for a group of columns within a table.
NonStop SQL/MX provides a method for generating histograms that shows how data is
distributed for a column within a table. The purpose of generating these statistics is to
enable the optimizer to create more efficient data query execution plans.
When generating a histogram for a table, NonStop SQL/MX divides the range of
specified column values of the table into some number of intervals, distributing the
rows evenly within these intervals. It computes statistics associated with each interval
and uses the statistics to devise optimized plans.
For further information, see the UPDATE STATISTICS Statement and User Metadata
(UMD)—Histogram Tables entries in the SQL/MX Reference Manual. The SQL/MX
Query Guide contains a section on keeping statistics current and best practices to
follow with respect to statistics.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

5 -9
Optimizer and Executor Statistics

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

5- 10
 6
Data Definition Language (DDL)
Differences
With NonStop SQL/MX Release 2, NonStop SQL/MX can now create objects. Prior to
Release 2, NonStop SQL/MX did not allow creation of objects. NonStop SQL/MX can
access NonStop SQL/MX as well as SQL/MP files. New Data Definition Language
(DDL) statements allow you to create and modify application files. For details on these
statements, see “SQL/MX Statements” in the SQL/MX Reference Manual.

SQL Statements
You can now create, alter, and drop SQL/MX database objects similarly to how you
manipulate these objects in NonStop SQL/MP, with these exceptions:
• You cannot create collations for SQL/MX tables, so there is no CREATE
COLLATION statement. You create constraints at CREATE TABLE time or later
with a ALTER TABLE ADD CONSTRAINT statement, so there is no separate
CREATE CONSTRAINT statement. You drop constraints with the ALTER TABLE
DROP CONSTRAINT statement.
• You cannot alter catalogs, collations, programs or views.
• The generic DROP command has been replaced by specific DROP CATALOG,
DROP INDEX, DROP PROCEDURE, DROP SCHEMA, DROP SQL, DROP
SQLMP ALIAS, DROP TABLE, DROP TRIGGER, and DROP VIEW statements.
See the SQL/MX Reference Manual for information about statements and database
objects.

Database Object Names

You must use ANSI logical names (a three-part logical name: catalog-name. schema-
name.object-name) to describe SQL/MX objects, using the LOCATION clause to
specify their physical location.
You can use physical names, ANSI logical names, SQL/MP alias names, or DEFINE
names to refer to SQL/MP objects.
See “SQL/MX Language Elements” in the SQL/MX Reference Manual for details about
database object names.

ANSI Names
Namespace for SQL/MX objects is organized in a hierarchical manner. Database
objects exist in schemas, which are themselves contained in catalogs. Catalogs are

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

6 -1
Data Definition Language (DDL) Differences Organization of Catalogs and Schemas

collections of schemas. Schema names must be unique within a given catalog. Multiple
objects with the same name can exist provided that each belongs to a different
namespace. NonStop SQL/MX supports various types of namespaces. See “SQL/MX
Language Elements” in the SQL/MX Reference Manual for details about these
namespaces.
You reference SQL/MP database objects by using their Guardian physical names. To
allow for the use of ANSI logical names, NonStop SQL/MX creates a metadata table,
the OBJECTS table. This table is used to store object names. Mappings from logical
object names to physical Guardian locations are stored in the metadata table
MP_PARTITIONS.
Use the CREATE SQLMP ALIAS command within your application to create needed
mappings from logical to physical names. This command inserts a mapping as a row in
the OBJECTS table. SQL/MP aliases are true ANSI names that represent the
underlying Guardian physical names of SQL/MP objects, in which the catalog part
refers to an existing catalog and the schema part refers to an existing schema.

Organization of Catalogs and Schemas

NonStop SQL/MX and NonStop SQL/MP have different concepts of catalogs.

SQL/MX Catalogs
An SQL/MX catalog is a named logical object that contains descriptions of a set of
schemas. You can access SQL/MX objects with the three-part name of the actual
object.
A catalog can contain multiple schemas, each possibly owned by a different user. A
catalog cannot contain other catalogs. Any user on a node can create a catalog on that
node. Any user on a node can drop an empty catalog on that node. For more
information, see “SQL/MX Language Elements” in the SQL/MX Reference Manual.
NonStop SQL/MX uses ANSI-compliant CREATE and DROP statements to create and
drop catalogs and schemas. Note that catalog names beginning with
“NONSTOP_SQLMX_” are reserved for system metadata.

SQL/MP Catalogs
An SQL/MP catalog is a set of tables and indexes that describe SQL objects. Tables in
the set are called catalog tables and SQL creates them, along with their indexes, when
you execute a CREATE CATALOG statement. Each SQL/MP catalog (the set of
catalog tables and their indexes) resides on its own Guardian subvolume, and the
name of that subvolume is also the name of the catalog.
Each node on which SQL/MP is used has one special catalog called the system
catalog and might have many other catalogs. Each table, view, index, partition,
collation, or catalog table located on a node must be described in a catalog on the

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

6 -2
Data Definition Language (DDL) Differences Metadata

same node. Normally, an SQL program is registered in a catalog, too. For more
information, see the SQL/MP Reference Manual.
NonStop SQL/MP allows you to create catalogs and to drop system catalogs.

Metadata
The system catalog is NONSTOP_SQLMX_nodename. SQL/MX metadata is
organized within these system schemas:
SYSTEM_SCHEMA System schema tables that list information such as
catalogs and schemas in the system.
DEFINITION_SCHEMA_ Definition schema tables that list information such as
VERSION_versnum columns in tables, constraints, partitions, views,
procedures and so on.
SYSTEM_DEFAULTS_ User-provided system defaults.
SCHEMA
MXCS_SCHEMA MXCS subsystem data source and governing
information.
SYSTEM_SQLJ_SCHEMA System schema that contains only the stored
procedure VALIDATEROUTINE, which is for internal
use.
SQL/MX’s metadata is on the same subvolume as the system catalog, which you
specify when you install NonStop SQL/MX.
SQL/MP metadata, also called the Data Dictionary, is organized by disk. It consists of
all the catalogs on a network and labels for each file on the volume. The label includes
the name or the object, the name of the catalog that describes the object, and other
information about the file.
SQL/MP metadata is the combination of system catalog metadata and user catalog
metadata. NonStop SQL/MP creates the system catalog when you issue the CREATE
SYSTEM CATALOG command, in whatever catalog you specify. The default is
$SYSTEM.SQL. Each catalog of user metadata resides on its own Guardian
subvolume.
Previous releases of NonStop SQL/MX use a different table scheme for metadata
where metadata resided in SQL/MP tables registered in the SQL/MP catalog. For
SQL/MX Release 2.0, metadata resides in SQL/MX metadata tables. The MIGRATE
utility is provided to move metadata from previous releases of NonStop SQL/MX to
current metadata tables. For more information about this utility, see the SQL/MX
Installation and Management Guide and the SQL/MX Reference Manual.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

6 -3
Data Definition Language (DDL) Differences Tables

Tables
You create constraints on an SQL/MX table with the CREATE TABLE and ALTER
TABLE ADD CONSTRAINT statements. You create constraints on SQL/MP tables with
the CREATE CONSTRAINT statement.

Clustering Keys
You assign a clustering key for an SQL/MX table by using the STORE BY clause of the
CREATE TABLE statement. If you enter STORE BY PRIMARY KEY, NonStop SQL/MX
bases the clustering key on the primary key. The STORE BY clause determines the
order of rows in the physical file and affects how you can partition the file.
You assign a clustering key for an SQL/MP table by using the CLUSTERING KEY
clause of the CREATE TABLE statement, listing the columns that will be used for this
key. SQL/MP stores rows in ascending or descending order, as you specify. You
cannot control the order of rows in the physical file.

Table Attributes
NonStop SQL/MX allows you to set these table attributes: ALLOCATE,
AUDITCOMPRESS, BLOCKSIZE, CLEARONPURGE, EXTENT, and MAXEXTENTS.
NonStop SQL/MP allows you to set these table attributes: ALLOCATE, AUDIT,
AUDITCOMPRESS, BLOCKSIZE, BUFFERED, CLEARONPURGE, DCOMPRESS,
EXTENT, FORMAT, ICOMPRESS, LOCKLENGTH, MAXEXTENTS, NOPURGEUNTIL,
RECLENGTH, TABLECODE, and VERIFIEDWRITES.

File Types
You can access these type of SQL/MP files through NonStop SQL/MX:
• Key-sequenced tables with or without partitions
• Entry-sequenced tables that are not partitioned
You cannot access these type of SQL/MP files through NonStop SQL/MX:
• Entry-sequenced tables that are partitioned
• Relative tables

Constraints
You create constraints on SQL/MX tables with the CREATE TABLE statement, and you
can add or drop constraints with the ALTER TABLE statement. For more information,
see “SQL/MX Language Elements” in the SQL/MX Reference Manual.
To create constraints on an SQL/MP table, use the SQL/MP CREATE CONSTRAINT
statement when you create the table. To drop constraints on an SQL/MP table, use the
SQL/MP DROP statement.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

6 -4
Data Definition Language (DDL) Differences Indexes

Indexes
You can create indexes for SQL/MX files, with these differences from NonStop
SQL/MP:
• Index table attribute
NonStop SQL/MX allows you to set these index table attributes: ALLOCATE,
AUDITCOMPRESS, BLOCKSIZE, CLEARONPURGE, EXTENT, and
MAXEXTENTS.
NonStop SQL/MP allows you to set these index table attributes: ALLOCATE,
AUDITCOMPRESS, BLOCKSIZE, BUFFERED, CLEARONPURGE,
DCOMPRESS, EXTENT, FORMAT, ICOMPRESS, ISLACK, LOCKLENGTH,
MAXEXTENTS, NOPURGEUNTIL, SERIALWRITES, SLACK,TABLECODE, and
VERIFIEDWRITES.
• Index table access
NonStop SQL/MX does not support WITH SHARED ACCESS.
• Partition changes
NonStop SQL/MX allows you to define secondary partitions as range or hash
partitions.
• Populating indexes
If you are creating an index on a large SQL/MX table that is already populated, you
should use the NO POPULATE option, and then run the POPULATE INDEX utility
to load the index. Because CREATE INDEX executes in a single TMF transaction,
it could experience TMF limitations such as a transaction timeout if a large amount
of data is to be moved. See the SQL/MX Installation and Management Guide for
information about creating and populating indexes and the SQL/MX Reference
Manual for details about the CREATE INDEX statement and the POPULATE
INDEX utility.
• PARALLEL EXECUTION ON | OFF option
NonStop SQL/MX uses the optimizer to determine whether to perform parallel
index loading. It is not a specifiable attribute.
• INVALIDATE | NO INVALIDATE option
SQL./MX does not allow you to control whether or not changes to indexes cause
similarity checks and possible recompiles.

Partitions
NonStop SQL/MX allows you to locate as many partitions as you want on the same
disk. Like SQL/MX tables, partitions use ANSI logical names. You can specify
Guardian names for a partition.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

6 -5
Data Definition Language (DDL) Differences Views

NonStop SQL/MP allows no more than one partition of an object on each disk.
SQL/MP partitions have Guardian names.
You manage partitions differently in NonStop SQL/MX and NonStop SQL/MP. For
details, see Partition Management: The MODIFY Utility on page 7-2

Views
SQL/MX Views
The distinction between protection and shorthand views does not exist for SQL/MX
views. To create a view, you must have SELECT privileges for the objects underlying
the view.
NonStop SQL/MX supports UNION in a view definition.
For information about SQL/MX views, see “SQL/MX Language Elements” in the
SQL/MX Reference Manual.

SQL/MP Views
An SQL/MP view is either a protection view or a shorthand view. A protection view is
derived from a single table and can be read, updated, and secured. A shorthand view
is derived from one or more tables or other views and inherits the security of the
underlying tables. A shorthand view can be read but not updated.
To create a view, you must have authority to write to the catalog that receives the view
description and to the USAGES tables of catalogs describing the underlying tables and
views.
To create a protection view, you must also be a generalized owner of the underlying
table. Any partitions or indexes of the table underlying the protection view must be
accessible when you create the view. To specify read or write access for a protection
view, you must have authority to read or write to the underlying table and all associated
indexes (unless you are the super ID).
NonStop SQL/MX and NonStop SQL/MP handle security on views differently. In
NonStop SQL/MP, you need to have appropriate security on the underlying tables of
the view in order to update the view. In NonStop SQL/MX you need only to have
appropriate security on view itself. You do not need to have privileges on the
underlying table.
For information about SQL/MP views, see Views in the SQL/MP Reference Manual.

Object Security
NonStop SQL/MX uses GRANT and REVOKE statements that conform to the ANSI
SQL standard to control privileges for objects. It does not support a SECURE clause.
Security is controlled internally by the software and does not use the security vector.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

6 -6
Data Definition Language (DDL) Differences Object Security

NonStop SQL/MP uses the SECURE clause to specify a Guardian security string,
“RWEP” to secure objects. security is controlled by the Guardian security vector stored
in the label.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

6 -7
Data Definition Language (DDL) Differences Object Security

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

6 -8
 7 Utility Differences
SQL/MX 2.0 includes utilities that perform tasks such as populating indexes, displaying
information about files or DDL, managing partitions, and so on. These SQL/MX utilities
are similar to those in NonStop SQL/MP, but not identical:

DISPLAY USE OF Command

In NonStop SQL/MX, DISPLAY USE OF is an MXCI command that provides usage
information only for statically compiled modules, such as:
• Modules using a specific SQL object such as a table, index, or MP alias
• All the SQL objects used by a module or set of modules
NonStop SQL/MP’s implementation of this command provides information on various
objects:
• All objects using a specific SQL object, such as a table, index, view, or collation.
NonStop SQL/MP displays all tables, views, indexes, collations, and programs
associated with that object.
• Optionally, detailed information on each object displayed.
For syntax and description of this command, see “MXCI Commands” in the SQL/MX
Reference Manual.

IMPORT Command
You can use SQL/MX’s IMPORT command to import data from a file (ASCII or binary)
into a NonStop SQL/MX table. IMPORT executes at the OSS or MXCI command
prompt. It is similar to but not equivalent to SQL/MP’s LOAD command. See “SQL/MX
Utilities” in the SQL/MX Reference Manual for details.

MXTOOL Utility
Earlier releases of NonStop SQL/MX supported only a limited set of tools with the
MXTOOL utility. NonStop SQL/MX Release 2 provides a tool, the MXTOOL utility, that
gives you additional options. The MXTOOl utility allows you verify the contents of an
SQL/MX object, fix parts of the metadata and label information, return information
about SQL/MX objects, and remove both damaged and undamaged objects from
metadata or labels in a uniform way.
The MXTOOL utility performs these operations:
FIXUP Similar to various SQL/MP operations

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

7 -1
Utility Differences Partition Management: The MODIFY Utility

GOAWAY Similar to SQL/MP’s GOAWAY command

INFO Displays ANSI names and versions of Guardian files used by an object
VERIFY Similar to SQL/MP’s VERIFY command
For more information, see “SQL/MX Utilities” in the SQL/MX Reference Manual.

Partition Management: The MODIFY Utility

In both NonStop SQL/MX and NonStop SQL/MP, you create partitioned tables and
indexes with CREATE TABLE and CREATE INDEX.
In NonStop SQL/MX you use the MODIFY utility to manage partitions. You can:
• Reuse an existing partition of a range-partitioned table
• Add, drop, or move partitions of range-partitioned tables and indexes
• Add, drop, or move partitions of hash-partitioned tables and indexes
• Move partitions of system-clustered objects
In NonStop SQL/MP, you use the PARTONLY MOVE clause of ALTER TABLE (for a
key-sequenced file) to break the table into partitions or to break a partition into
additional partitions, or the ADD PARTITION clause of ALTER TABLE to add a partition
to the end of an entry-sequenced or relative table. To create a partitioned index, use
CREATE INDEX or the PARTONLY MOVE clause of ALTER INDEX.
For a description of the MODIFY command, see “SQL/MX Utilities” in the SQL/MX
Reference Manual.

POPULATE INDEX Command

NonStop SQL/MX includes a CREATE INDEX command, but it includes a NO
POPULATE option you can use if you prefer to populate the index in a separate
operation from creating it. For instance, if you are creating an index on a large table
that is already populated, you should use the NO POPULATE option and then run the
POPULATE INDEX utility to load the index. Because CREATE INDEX executes in a
single TMF transaction, it could experience TMF limitations such as a transaction
timeout if there is a large amount of data to be moved.
The SQL/MX statements CREATE INDEX NO POPULATE followed by POPULATE
INDEX are equivalent to the SQL/MP CREATE INDEX statement.
For more information, see the SQL/MX Reference Manual.

SHOWDDL Command
SQL/MX’s SHOWDDL command displays a DDL statement that would create a table,
view, or stored procedure as it exists in metadata, including the object’s dependent

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

7 -2
Utility Differences SHOWLABEL Command

objects. This output can be used to re-create the specified object, including its
dependent objects.
SHOWDDL is similar to SQL/MP’s INVOKE command, which produces a record
description that corresponds to a row in a specified table or view.
See “MXCI Commands” in the SQL/MX Reference Manual for more information.

SHOWLABEL Command
SQL/MX’s SHOWLABEL command displays file-label and resource-fork information for
SQL/MX tables, worktables, views, and indexes. The information includes the object
version, physical location, and other characteristics.
Every SQL object includes a logical file label to store the object’s file attributes and
information about its dependent objects. The resource fork is a new file that contains
structural descriptions of a table. When an SQL/MX object is created, two physical files
are instantiated: the data fork and the resource fork. The data fork is where the user
data resides. The resource fork contains structural information, such as the partition
map.
SHOWLABEL is similar to SQL/MP’s FILEINFO utility, which displays information
about the versions and physical characteristics of tables, indexes, views, collations,
Enscribe files, and OSS files.
For more information, see “MXCI Commands” in the SQL/MX Reference Manual.

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

7 -3
Utility Differences SHOWLABEL Command

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

7 -4
 Index
A B
Access options BACKUP utility 2-2
differences 3-16 BASETABS table 2-2
STABLE access 3-17 BEGIN DECLARE SECTION statement 2-2
summary 2-1 BEGIN WORK statement 2-2, 3-18
Access plans BETWEEN predicate 2-2
forcing 5-8 BLOCKSIZE file attribute 2-2
influencing 5-7 BREAK FOOTING command 2-2
reviewing with EXPLAIN function 5-8 BREAK ON command 2-2
BREAK TITLE command 2-2
ADD DEFINE 2-1
BUFFERED file attribute 2-2
Aggregate functions 2-1, 3-11
Aliases 2-1
ALLOCATE file attribute 2-1 C
ALTER CATALOG statement 2-1 c89 2-5
ALTER COLLATION statement 2-1 Cache 3-24/3-26
ALTER DEFINE statement 2-1 CANCEL command 2-2
ALTER INDEX statement 2-1 CASE expression 2-2
ALTER PROGRAM statement 2-1 CAST function
ALTER TABLE statement 2-1 and descriptor area 4-8
ALTER VIEW statement 2-1 data conversion 4-3
ANSI compatibility 1-5 CATALOG command 2-2
APPEND command 2-1 Catalogs 6-2
APPENDCANCEL command 2-1 CATALOGS table 2-3
APPENDRESTART command 2-2 CENTER_REPORT option 2-3
AS clause 2-2 Character data type 2-3
AS DATE/TIME clause 2-2 Character expression 2-3
ASCII character set 2-2 Character sets 2-3
AUDIT file attribute 2-2 CHAR_LENGTH function 2-3
AUDITCOMPRESS file attribute 2-2 CLEANUP command 2-3
Audited tables 2-2 CLEARONPURGE file attribute 2-3
AUTOCOMMIT CLOSE statement 2-3
example of 3-20 Clustering keys 2-3
for transactions 3-19 COLLATE clause 2-3
AVG function 2-2 Collation definitions 2-3
Collations 2-3
Column Identifier 2-3
Columns 2-3
COLUMNS table 2-3

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

Index -1
Index D

COMMENT statement 2-3

Comments 2-3
D
COMMENTS table 2-3 Data Definition Language statements 6-1
COMMIT option 2-3 Data dictionary 2-5
COMMIT WORK statement 3-19 Data types
Comparison predicate 2-3 conversion 4-3
Compile date-time 3-1
SQL/MP 4-10 description of 3-1
SQL/MX 4-11 differences 3-1
COMPUTE_TIMESTAMP function 2-3 interval 3-2
CONCAT clause 2-3 VARCHAR 4-2
Constraints 2-4, 6-4 Database object names 6-1
CONSTRNT table 2-4 Data-flow task model 5-5
CONTINUE statement 2-4 DATE data type 2-5
CONTROL EXECUTOR directive 2-4 DATEFORMAT function 2-5
CONTROL QUERY directive 3-23 DATETIME data type 2-5
CONTROL TABLE directive 3-21 Datetime data types 3-1
Conversion, data types 4-3 Datetime literals 3-6
CONVERT command 2-4 Datetime value functions 3-9
CONVERTTIMESTAMP function 2-4 Date-time data types 2-5
COPY command 2-4 Date-time functions 2-5
Correlation names 2-4 Date-time literals 2-5
COUNT function 2-4 DATE_FORMAT option 2-5
CPRLSRCE table 2-4 DAYOFWEEK function 2-5
CPRULES table 2-4 DCL (Data Control Language)
CREATE CATALOG statement 2-4 statements 2-6
CREATE COLLATION statement 2-4 DCOMPRESS 2-6
CREATE CONSTRAINT statement 2-5 DEALLOCATE file attribute 2-6
CREATE INDEX statement 2-5 DECIMAL_POINT option 2-6
CREATE SYSTEM CATALOG DECLARE CURSOR statement 2-6
statement 2-5 DEFAULT clause 2-6
CREATE TABLE statement 2-5 Default settings
CREATE VIEW statement 2-5 SQL/MP 3-23
CURRENT function 2-5 SQL/MX 5-7
CURRENT_TIMESTAMP function 2-5 DEFINEs 1-2
Cursors DELETE DEFINE statement 2-6
summary 2-5 DELETE statement 2-6
updatable 4-6 Delimited identifiers 3-5
DESCRIBE INPUT statement 2-6
DESCRIBE statement 2-6
Descriptor area 4-7

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

Index -2
Index E

DETAIL command 2-6 Expressions 2-7

Diagnostics area 4-9 Expressions, interval value 3-10
Differences, NonStop SQL/MP and EXTEND function 2-8
NonStop SQL/MX 1-2 EXTENT file attribute 2-8
DISPLAY STATISTICS command 2-6
DISPLAY USE OF command 7-1
DISTINCT clause 2-6
F
DML features 3-1 FC command 2-8
DML statements 3-12 FETCH statement 4-6
DOWNGRADE CATALOG command 2-6 File attributes 2-8
DOWNGRADE SYSTEM CATALOG File organization 2-8
command 2-7 File types 3-4
DROP statement 2-7 FILEINFO command 2-8
DROP SYSTEM CATALOG statement 2-7 FILENAMES command 2-8
DSL (Data Status Language) FILES command 2-8
statements 2-7 FILES table 2-8
DSLACK file attribute 2-7 Filesets 2-8
DUP command 2-7 FIXUP command 7-1
Dynamic SQL 4-6 Floating point data type 2-8
FREE RESOURCES statement 2-8
E Functions
aggregate 3-11
EDIT command 2-7
date-time value 3-9
Embedded SQL 4-1
END DECLARE SECTION statement 2-7 mathematical 3-10
ENV command 2-7 numeric value 3-9
ERROR command 2-7 sequence 3-10
Error handling 4-5, 4-8, 4-9 SETSCALE 4-1
EXECUTE IMMEDIATE statement 2-7 FUP command 2-8
EXECUTE statement 2-7
Executor G
design differences 5-3
Generalized owner 2-8
query plan 5-5 GET CATALOG OF SYSTEM
queues 5-5 statement 2-8
SQL/MP iterator model 5-4 GET DESCRIPTOR statement 1-2
SQL/MX task model 5-5 GET DIAGNOSTICS statement
EXISTS predicate 2-7 C example 4-10
EXIT command 2-7 description of 4-9
EXPLAIN directive 2-7 GET VERSION OF PROGRAM
EXPLAIN function, reviewing access statement 2-8
plans 5-8 GET VERSION statement 2-8
Explicit transactions 3-18 GOAWAY command 7-2

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

Index -3
Index H

Group manager 2-8 INTERVAL literal 2-10

Guardian names 2-8 Interval literals 3-8
Interval value expressions 3-10
H INVALIDATE option, indexes 6-5
INVOKE directive 4-3
HEADING clause 2-8
ISLACK file attribute 2-10
HEADINGS option 2-8
Isolation level
HELP command 2-9
differences 3-16
HELP TEXT statement 2-9
STABLE access 3-17
Histograms 5-9
statement level 3-16
HISTORY command 2-9
Host identifiers 2-9 transaction level 3-16
Host programs 2-9
Host variables J
declaring 4-1 Joins 2-10
summary 2-9 Joins, types of 3-15
JULIANTIMESTAMP function 2-10
I
ICOMPRESS file attribute 2-9 K
Identifiers 3-5 Keys 2-10
IF/THEN/ELSE clause 2-9 KEYS table 2-10
IMPORT command 7-1
IN predicate 2-9
INCLUDE SQLCA directive 2-9
L
LEFT_MARGIN option 2-10
INCLUDE SQLDA directive 2-9
LIKE predicate 2-10
INCLUDE SQLSA directive 2-9
Limits 2-10
INCLUDE STRUCTURES directive 2-9
Index keys 2-9 LINE_NUMBER function 2-10
LINE_SPACING option 2-10
Indexes
description of 6-5 LIST command 2-10
Literals
populating 7-2
date-time 3-6
INDEXES table 2-9
string delimiters 3-5
Indicator parameters 2-9
LOAD command 2-10
Indicator variables 4-3
LOCK TABLE statement 2-10
INFO command 7-2
Locking 2-10
INFO DEFINE statement 2-9
LOCKLENGTH file attribute 2-10
Inherited transactions 3-19
Locks, releasing 4-5
INITIALIZE SQL command 2-9
LOG command 2-10
INSERT statement 3-14
LOGICAL_FOLDING option 2-10
INTERVAL data type 2-10
Interval data type 3-2

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

Index -4
Index M

M Object namespaces 6-1

OCTET_LENGTH function 2-11
Mathematical functions 3-10 OPEN statement 2-11
MAX function 2-11 Open System Services (OSS) 1-3
MAXEXTENTS file attribute 2-11 Optimizer, design differences 5-1
Message file 2-11 OSS names 2-11
Metadata 6-3 OUT command 2-11
MIN function 2-11 OUT_REPORT command 2-12
MODIFY CATALOG command 2-11 OVERFLOW_CHAR command 2-12
MODIFY LABEL command 2-11 OWNER file attribute 2-12
MODIFY REGISTER command 2-11
Multibyte character sets 2-11
MXTOOL utility 7-1
P
PAGE TITLE option 2-12
N PAGE_COUNT option 2-12
PAGE_FOOTING option 2-12
NAME command 2-11 PAGE_LENGTH option 2-12
NAME option 2-11 PAGE_NUMBER option 2-12
Name resolution 1-3 PARALLEL EXECUTION, index
Names 2-11, 6-1 loading 6-5
Namespace 6-1 Parallel index loading 2-12
NEWLINE_CHAR option 2-11 Parameters 2-12
NO INVALIDATE option, indexes 6-5 PARTITION clause 2-12
Nonaudited tables 2-11 Partition management 2-12
NonStop SQL/MP 1-1 Partitions 6-5
NonStop SQL/MX 1-1 PARTNS table 2-12
description of 1-1 Performance 1-4
parallel operations 1-4 PERUSE command 2-12
questions about differences 1-2 Plans 2-12
redesigned optimizer and executor 1-4 Populating indexes 7-2
NOPURGEUNTIL file attribute 2-11 POSITION function 2-12
NULL predicate 2-11 Positioned UPDATE Statement 4-6
Null value 2-11 Predicates
NULL_DISPLAY 2-11 LIKE 3-11
Numeric data types 2-11 row-value constructors 3-11
Numeric literals 2-11 PREPARE statement 2-12
Numeric value expression and Primary keys 2-12
functions 3-9 Print item 2-12
PROGID file attribute 2-12
O Program compilation
OBEY command 2-11 SQL/MP 4-10
Object names 6-1 SQL/MX 4-11

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

Index -5
Index Q

Program invalidation 2-13 Report Writer commands (continued)

PROGRAMS table 2-13 CANCEL command 2-2
Protection view 2-13 CENTER_REPORT option 2-3
PURGE command 2-13 CONCAT clause 2-3
PURGEDATA command 2-13 DATE_FORMAT option 2-5
DECIMAL_POINT option 2-6
Q DETAIL command 2-6
Qualified fileset list 2-13 HEADINGS option 2-8
Quantified predicate 2-13 IF/THEN/ELSE clause 2-9
Query caching LEFT_MARGIN option 2-10
cache location 3-24 LINE_NUMBER function 2-10
considerations 3-24 LINE_SPACING option 2-10
controls 3-24 LIST command 2-10
counter 3-26 LOGICAL_FOLDING option 2-10
description of 3-24 NAME command 2-11
efficiency increase 3-24 NEWLINE_CHAR option 2-11
setting 3-24 NULL_DISPLAY 2-11
size limit 3-24 OUT_REPORT command 2-12
statistics 3-25 OVERFLOW_CHAR command 2-12
types of queries 3-26 PAGE TITLE option 2-12
Query expressions 3-15 PAGE_COUNT option 2-12
Query type comparison 3-27 PAGE_FOOTING option 2-12
Queues 5-5 PAGE_LENGTH option 2-12
PAGE_NUMBER option 2-12
R REPORT FOOTING command 2-13
READ UNCOMMITTED access, compared REPORT TITLE command 2-13
to BROWSE access 3-16 RESET LAYOUT command 2-13
RECLENGTH file attribute 2-13 RESET REPORT command 2-13
Regular identifiers 3-5 RESET SESSION command 2-13
RELEASE statement 2-13 RESET STYLE command 2-13
REPORT FOOTING command 2-13 RIGHT_MARGIN option 2-13
REPORT option 2-13
ROWCOUNT option 2-13
REPORT TITLE command 2-13
SET LAYOUT command 2-14
Report Writer 1-3
SET PARAM command 2-14
Report Writer commands
SET STYLE command 2-14
AS clause 2-2
SHOW LAYOUT command 2-14
AS DATE/TIME clause 2-2
SHOW REPORT command 2-14
BREAK FOOTING command 2-2
SHOW SESSION command 2-14
BREAK ON command 2-2
SHOW STYLE command 2-14
BREAK TITLE command 2-2

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

Index -6
Index S

Report Writer commands (continued) Set functions

SPACE option 2-15 See Aggregate functions
SUBTOTAL command 2-16 SET LAYOUT command 2-14
SUBTOTAL_LABEL option 2-16 SET PARAM command 2-14
TIME_FORMAT option 2-16 SET SESSION command 2-14
UNDERLINE_CHAR option 2-17 SET STYLE command 2-14
VARCHAR_WIDTH option 2-17 SETSCALE function 4-1
WINDOW option 2-18 Shorthand view 2-14
SHOW CONTROL command 2-14
Reserved words 3-5
SHOW DEFINE statement 2-14
RESET DEFINE statement 2-13
SHOW DEFMODE command 2-14
RESET LAYOUT command 2-13
SHOW LAYOUT command 2-14
RESET PARAM command 2-13
SHOW PARAM command 2-14
RESET PREPARED command 2-13
SHOW PREPARED command 2-14
RESET REPORT command 2-13
SHOW REPORT command 2-14
RESET SESSION command 2-13
SHOW SESSION command 2-14
RESET STYLE command 2-13
SHOW STYLE command 2-14
RESETBROKEN file attribute 2-13
SHOWDDL command 7-2
RIGHT_MARGIN option 2-13
SHOWLABEL command 7-3
ROLLBACK WORK statement 3-19
Similarity checking 1-3, 2-15
ROWCOUNT option 2-13
SLACK file attribute 2-15
SMF 2-16
S SPACE option 2-15
Sample database 2-13 SQL directive 2-15
SAVE command 2-14 SQL identifiers 2-15
Scheduler-driven task model 5-5 SQL statements, DML differences 3-12
Schemas 6-2 SQLCA 1-2, 4-5, 4-9
Search conditions 2-14 SQLCI 2-15
SECURE command 2-14 SQLCI commands 2-15
SECURE file attribute 2-14 SQLCODE 4-8
SELECT statement SQLCOMP command 2-15
placement of GROUP BY clause 3-15 SQLDA 4-6
summary of 2-14 SQLDA area 2-15
table references 3-15 SQLSA 4-10
Sequence functions 3-10 SQLSA area 2-15
SERIALIZABLE access and REPEATABLE SQLSTATE 4-8
access 3-17 SQL/MP
SERIALWRITES file attribute 2-14 See NonStop SQL/MP
SET DEFINE statement 2-14 SQL/MP catalogs 6-2
SET DEFMODE command 2-14 SQL/MX
SET DESCRIPTOR statement 1-2, 4-7 See NonStop SQL/MX

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

Index -7
Index T

SQL/MX catalogs 6-2 TACL 1-3

STABLE access TEDIT command 2-16
description of 3-17 Temporary tables 2-16
EXCLUSIVE MODE 3-18 TIME data type 2-17
READ COMMITTED 3-17 TIMESTAMP data type 2-17
SQL/MX compared to SQL/MP 3-17 TIME_FORMAT option 2-16
Standards conformance 2-15 TMF transactions 2-17
Statements TOTAL command 2-17
DDL (Data Definition Language) 6-1 Transaction management
DML (Data Manipulation auto abort 3-19
Language) 3-12 AUTOCOMMIT 3-19
DSL (Data Status Language) 2-7 cursor operations 3-21
Statistics ending a transaction 3-18
SQLSA 4-10 example of 3-19
updating 5-9 explicit transactions 3-18
Stored procedures 2-16 implicit transactions 3-18
String functions 2-16 inherited transactions 3-19
String literals 2-16 SET TRANSACTION statement 3-19
Subqueries 2-16 starting a transaction 3-18
SUBSTRING function 2-16 system-defined 3-18
SUBTOTAL command 2-16 user-defined transactions 3-18
SUBTOTAL_LABEL option 2-16 Transaction Management Facility
SUM function 2-16 (TMF) 3-18
Super ID 2-16 TRANSID table 2-17
SYSKEY 2-16 TRIM function 2-17
System catalog 2-16
SYSTEM command 2-16
System DEFINEs 2-16
U
UNDERLINE_CHAR option 2-17
SYSTEM_DEFAULTS table 5-7
UNLOCK TABLE statement 2-17
Updatable cursors 4-6
T UPDATE statement 3-16
Table references 3-15 UPDATE STATISTICS statement 2-17
TABLECODE file attribute 2-16 UPGRADE CATALOG command 2-17
Tables UPGRADE SYSTEM CATALOG
attributes of 6-4 command 2-17
clustering keys 6-4 UPSHIFT function 2-17
constraints 6-4 USAGES table 2-17
description of 6-4 User-defined transaction 3-18
file types 6-4 Utilities 7-1
TABLES table 2-16

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

Index -8
Index V

V
VARCHAR data type 4-2
VARCHAR_WIDTH option 2-17
VERIFIEDWRITES file attribute 2-17
VERIFY command 7-2
Versioning 3-29
Versions 2-18
VERSIONS table 2-18
Views 6-6
VIEWS table 2-18
VOLUME command 2-18

W
WHENEVER declarative 4-9
WHENEVER directive 2-18
WHERE clause 2-18
WINDOW option 2-18
WITH SHARED ACCESS option 2-18

Special Characters
! command 2-18

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

Index -9
Index Special Characters

HP NonStop SQL/MX Comparison Guide for SQL/MP Users—523735-003

Index -10