Business Objects Tips and Tricks

· Business Objects at Penn

The following are tips and tricks for using Business Objects Desktop Intelligence more
effectively, and for working around commonly experienced errors. If you have questions
which are not answered by the tips below, or have suggestions for additions to the list,
please contact Data Administration.

· Retrieving Data
· Saving and Exporting Queries and Data
· Report Formatting
· Password issues
· Installation issues
· Other error messages

Retrieving Data

• When using the "in list" feature in Deski, if you are typing items in yourself
(rather than picking them from a list of values), the items in your list should be
separated by commas. Do not type in any quote marks -- Desktop Intelligence will
add them itself -- and do not put a space after a comma. However, if you are
filling in a list of values in response to a prompt, you should separate the values
with a semicolon, rather than a comma (but still no space between values).
• If you notice that the objects in your universe aren't refreshed, be sure that
when you first log on to Desktop Intelligence that the box labeled "Use is Offline
Mode" is not checked.

• If you get a "No data to fetch" message, and you're pretty sure you should be
getting something, check to make sure you have configured your conditions
correctly. For example: everything in the student data collection uses upper case
letters. Another example: if you are entering a condition for a date in any of the
financial data collections, use the full year, in other words, 06-01-2008 (instead of
06-01-08)

• If you get a "Table or View does not exist" message, check to make sure the
data collection is available on the warehouse status page.

• To determine how many rows your query retrieved, choose "View" from the
Data menu, then click on the Definition tab of the Data Manager window. The
most recent execution of the query will be listed first, by date and time the query
was executed, along with the number of rows retrieved, and the amount of time
the database took to execute the query (please note that this time is not the time
elapsed on your desktop machine).
• When first testing a query you may wish to limit the number of rows retrieved
to determine if your results are as expected. You can do this by clicking the
Options button at the lower left corner of the Query Panel. Select "10 Rows", "20
Rows" or enter another value by clicking "Other" in the Partial Results area of the
window. Remember to return the setting to "Default Value" when you are ready
to retrieve all rows matching your query criteria.

• You can use use wildcards in conjunction with the "Matches Pattern"
operator to retrieve data that is like a value, rather than exactly equal to it. For
example, using a percent sign ( % ) in the condition COA_Fund Matches Pattern
5% will retrieve all funds (to which you have access) from 500000-599999.
Similarly, if you're trying to match almost the exact syntax, you can use an
underscore ( _ ). For example, COA_ORG Matches Pattern '91_2' will retrieve
ORG values 9132, 9142 and 9152.

• If you have multiple queries in one report (one .rep file), you may want to
rename your queries to better describe their use, so that when you go to edit or
refresh them, you'll have a better idea of what data you'll be retrieving. To do this,
go to the Data menu, and select "View Data" to display the Data Manager
window. The General section of the Definition tab contains a field called "Name",
which you can reuse to rename the query from the default "Query x with
Universe" (i.e., "Query 1 with FINQUERY").

• If you are trying to create a User Defined Object (UDO), but the universe you
are in won't let you (in the Query Panel, clicking on User Objects does nothing, or
Deski suddenly quits) try this:
Log out, delete the .udo file for that universe, log back in and try creating the User
Objects again. (Caution! doing this will remove any user objects you previously
created for this universe; you will have to re-create them.)
User Defined Objects all live on your local computer. They reside within one file
per universe, in the following path:
C:\Documents and Settings\<user_name>\Application Data\Business
Objects\Business Objects 11.5\Universes\
(where <user_name> is your user name on your computer)
The file names will be <universe_name>.udo
(where <universe_name> is the name of the universe, for example,
"FINQUERY.udo" holds your local User Defined Objects for the FINQUERY
universe.)

• You can re-use queries between similar universes, so long as all the result
objects and conditions from your query exist in the universe to which you'd like
change. To do this, go to the Data menu, and select "View Data" to display the
Data Manager window. The General section of the Definition tab contains a field
called "Universe", which has a small button with "..." directly to the right. When
you first display this window, the Universe field will display the universe against
which the query is currently directed. Click the "..." button to display a list of all
other universes to which you have access. Choose the one to switch your query to,
 and click ok. As long as all the objects in your query are available in the new
target universe, that universe name will now appear in the Universe field. (This is
convenient for switching from the FINQUERY to FINQUERY Template
Universe, for example.)

Saving and Exporting Queries and Data

• If you have taken the time to create a query, but want to be able to save your
work without running it, you have two options. The most convenient is to use
click the "Save and Close" button in the Query Panel to simply save the structure
of your report. If you do this, be sure to then use "Save As" to save it with a
distinctive name, in the location you wish. You can then open the query and
simply click the Refresh button to retrieve your data. Alternatively, once you've
built your query, you can click the Options button at the bottom left of the Query
Panel, and click the "Do Not Retrieve Data" checkbox, click OK, and then click
Run. Again, only the structure of your report will be displayed, which you can
then save. Keep in mind, though, that you must uncheck the Do Not Retrieve Data
checkbox before you choose to run the query in the future.

• By default, Desktop Intelligence saves queries (.rep files) in \My

Documents\My Business Objects Documents\userDocs. If you work with
several universes, you may find it convenient to create subfolders within that
directory to better organize your work.

• If you want to export the data you've retrieved for use in another application,
you have a couple options:

Save As: Desktop Intelligence allows you to use the "Save As" feature to save
documents in Excel, Adobe Acrobat PDF or CSV formats. Once you've refreshed
a report, click on the File menu and select Save As. Select the file format you'd
like from the "Save as type" drop-down list. Desktop Intelligence will then save
your file with the data as it appears on the screen. This means the output will be
saved and reflect section breaks, filters, special formatting, etc. The results of any
variables or calculations will be saved as text, not underlying formulae. If your
report has multiple tabs, and you choose to save as Excel, each tab will appear as
a separate worksheet within one workbook. Similarly, multiple tabs will be
accessible indidually in Adobe Acrobat (PDF) files. InfoView users can perform a
similar action. If you experience undesired changes in Excel color formatting
when overwritting a Excel file previoulsy saved from Deski, save it instead to
different file name.
(Note: This feature was not available in Business Objects 5.1.4 - the work around
was to use the Business Objects Edit menu and select "Copy All", then open a
 blank worksheet in Excel and choose "Paste Special" from the Excel Edit menu.
Choose the "Unformatted Text" option and click Ok to paste your data into the
worksheet.)

Export: You can also export the raw data retrieved in your query (retaining no Business
Objects post-query formatting or calculations), in a variety of file formats. From the Data
menu, select "View Data" to display the Data Manager window. Click the Export button
at the bottom of the Results tab, and designate file name and path, and file format
(options include text formats, and .xls for Excel). If you choose the All files (*.*) format,
you can also control the field delimiters used. [A word of caution about using the .xls
format: some versions of Excel object to the field names used in the Business Objects
classes. If you encounter an error when exporting/importing using the Excel format, try it
again with either the All Files or the Text Files format, and then use the Excel Import
Wizard to bring the data into your speadsheet.]

Report Formatting

• If you're having a problem displaying or printing pages after the first page of
a report, and you have an image of some sort as part of your report layout, check
to see whether the image is part of a table with other elements. If it is, move it
outside the table and see if this fixes your problem.

• To minimize the processing time for your query, consider using the Desktop
Intelligence toolbar icons or the Slice and Dice panel (rather than the query itself)
to perform the following operations on the desktop once your query has finished:

Sort the data: You can place sorts on multiple objects using the Slice and Dice
panel. You may also add breaks or report sections based on objects, and then sort
those as well.

Filter the data: Once you have retrieved all the data meeting your conditions you
may wish to further filter the data. Filtering via the Slice and Dice Panel enables
you to include or exclude specific values retrieved. This is particularly useful
when working with a large data set that you wish to manipulate many ways before
deciding upon the final report format. You may set, change and remove filters as
often as you wish without having to rerun the original query.

Perform calculations: You may add calculations to result objects, such as Sum,
Count, Average, Minimum, Maximum and Percentage. This allows you to both
view the data at the level of detail retrieved by your query and calculate at break
levels you have designated. Available calculations are dependent upon the
datatype of the object. More advanced calculations are also available using report
Variables.
 • Duplicate rows - to show or not to show

Unless you intentionally go to the Options in the Query Panel and select “no
duplicate rows”, Business Objects will return duplicate rows if they exist in the
database.
However, Desktop Intelligence will by default show the aggregate of those rows
in the report. In other words, the report may not be showing you all the underlying
data. You can see all the rows if you go to the “View Data” tool, but if everything
on two rows is identical in every column, then Deski will only display one row in
the resulting table on the report.

If this isn’t what you want it to do, to force Deski to show all rows, do the
following:

o Right click on the table in your report, and select Format Table.
o On the General tab, check the box that says “Avoid Duplicate Rows
Aggregation” and then click OK.

Password Issues

Desktop Intelligence and InfoView do check your password when you log in. However,
you must be sure to use the password change application to synchronize your Data
Warehouse, Business Objects and Business Objects DB Credentials each time you
change your Data Warehouse password. (The same applies if you are using Business
Objects to query other Oracle databases, such as Penn Community.) If your passwords
are not synchronized, your query attempts will return errors:

o If you check only the Business Objects boxes in the PassWord Changer
application, and not the Data Warehouse and other database boxes, when
you submit a query you will get a message saying your access is denied
for password errors.
o If you check the database boxes (i.e., Data Warehouse) and only one of the
Business Objects boxes, you will get this error:
Connection or SQL sentence error: (DA0005)
Exception: CS, Unexpected behavior

The resolution for all of the above situations is to go to the password change application
and reset your password, and be sure to check ALL of the boxes.

Installation Issues
 • Symptoms are: You retrieve the installation files and unzip them, double-click the
installer.bat, and get a message: "Please go to the control panel to install and
configure system components." This can happen if you downloaded the
installation files to a drive that is different from the one on which you are trying to
install the BusinessObjects Deski client. Try downloading and unzipping to the
user's desktop, and run the installer.bat from there.
• Symptoms are: You either 1) get a DA0005 error, and when you click on the
error message "details" it says that the DBDriver failed to load, or 2) you are
installing Deski on a remote computer and you get an "openSessionLogon" error
telling you it cannot establish a CMS connection.
Resolution: You must log in to InfoView to activate Desktop Intelligence the first
time you use it following installation. See the "Configure the Client" step in the
Installation instructions for details.
• You get an "Operation TimeOut" error and the message that "Your internet server
is not responsive." This probably means you are trying to use an older version of
Business Objects (for example, this can happen when your desktop shortcut is still
pointing to the old version). When you launch, make sure you are actually using
Business Objects Enterprise XI release 2 / Desktop Intelligence.

Other Error messages and error conditions

• Reports with date prompts returning no data: If you're not getting data back
when running reports with date prompts (and the same values worked in Business
Objects 5), make sure you're entering the date in 4-digit year format. For example,
rather than entering "7/1/07" (for July 1, 2007), enter "7/1/2007".
• If you attempt to open a report and you get an error:
You are not authorized to use this document. (FRM0008)
try one of the following solutions:
1) Is this a Corporate Document? In which case, do not try to open the local copy,
but instead go to File->Import from Repository...
and use the new one from the repository.
2) If this is a document you created yourself on your local computer using the old
version (or modified and saved on your local computer using the old version),
then, if possible, use the old version -- version 5.1.4 -- to open it. Go to File-
>Save As... and before you save it, click the box next to "Save for all users" in the
lower left corner of the screen. Then click on Save, and close it. You should now
be able to open it using Deski.
3) You can also get this error if you created this document yourself on your local
computer, then exported it to the Repository (for example, to your "Favorites" box
in InfoView), and then subsequently deleted the one on the Repository. The local
one will no longer be available to you -- or anyone else! To avoid this problem,
always make a backup copy of your local document, before you delete a
report on the Repository.
• If you've run a query, but are not seeing any results on screen other than the
column header cells, you may want to check to see if any of the following
features are set:

·You are viewing the Structure of the report, rather than the actual data retrieved.
(Check the View menu, Strucure setting.)
·There are Filters in your report. (Click on any cell that you can see, and choose
Filters from the Format menu.)
·The report is folded, so that only headers appear. (First check the old checkbox
on the General tab of the Standard Report Styles window, accessible from the
Tools menu. If Fold there is unchecked, you may also want to check the same
setting in Format/Table, to see if the report is folded there.

• You log in to Deski and try to refresh a report, but get: "Error During SQL
Execution: (DA0003) - CS, Job already in use".
One possible cause of this error message is because Internet Explorer 7 has a
default time out of 30 seconds compared with the Internet Explorer 6 time out of
60 seconds. For some queries, this is not enough time to refresh the entire query,
since the client is connecting using an HTTP protocol to the server.
It is necessary to extend the time-out limit on Internet Explorer 7. This can be
done by adding a key to the registry.
(CAUTION! The following resolution involves editing the registry. Using the
Registry Editor incorrectly can cause serious problems that may require you to
reinstall the Microsoft Windows operating system. Use the Registry Editor at your
own risk. It is strongly recommended that you make a backup copy of the registry
files before you edit the registry. End users should NOT attempt this without first
consulting with their Local Support Providers.)

To resolve the error message (if you are using IE 7)

1. Click Start > Run. The Run dialog box appears.
2. Type “regedit” in the Open: text box. Click OK. The Registry Editor appears.
3. Navigate to
HKEY_CURRENT_USER\Software\Microsoft\Windows\CurrentVersion\Interne
t Settings.
4. Add a ReceiveTimeout DWORD value with a data value of (<number of
seconds>)*1000. For example, if the required time out duration is eight minutes,
set the ReceiveTimeout data value to 480000 (<480>*1000).
5. Exit the Registry Editor.
6. Restart the computer.

• If your Business Objects menu bar disappears, try the following steps:
1. From the Windows 'Start' menu, select 'Run'.
2. Type "regedit" in the 'Run' dialog box.
 3. Locate and delete this folder in the registry:

HKEY_CURRENT_USER\Software\Business Objects\Suite
11.5\default\BusinessObjects\Application
Preferences\BusinessReporter\Desktop Intelligence
4. Restart Desktop Intelligence
5. If the steps above do not resolve the issue, try deleting the Desktop
Intelligence folder under HKEY_USERS & the user's SID:

HKEY_USERS\<user's SID>\ Software\Business Objects\Suite

11.5\default\BusinessObjects\Application
Preferences\BusinessReporter\Desktop Intelligence

6. Restart Desktop Intelligence

• If the Classes and Object pane disappears from the Query Panel, try the
following steps:
1. From the Windows 'Start' menu, select 'Run'.
2. Type "regedit" in the 'Run' dialog box.
3. Locate and delete these two registry keys:
 HKEY_CURRENT_USER\Software\Business Objects\Suite
11.5\default\BusinessObjects\Administrator User Prefs
 HKEY_CURRENT_USER\Software\Business Objects\Suite
11.5\default\BusinessObjects\Application
Preferences\BusinessDesigner
4. Restart Desktop Intelligence

• If you're having problems accessing the Slice and Dice window, click the Slice
and Dice button again, and, when the window doesn't appear, hit Alt+space, and
then Maximize the window.
Components of a Report
Within a given report, you have several components: sections, blocks, and cells. As you
explore information within a report, these components are not particularly important; as
you try to format the report or add information to it, however, it’s important to
understand which component you are altering.

Sections
Every report has a main section. Within the main section, you can have a section header
and a section footer; these are different from page headers and footers that appear in
printed reports. Main section headers typically hold the title of the report but also may
contain a picture or logo. Reports also may have subsections if you create a Master Detail
report. Figure shows a report with section headers for each different color of wine. The
report title “Average Wine Ratings By Country and Decade” appears in the main section.

Blocks
A block is a set of data that contains column headings, row headings, and data values. A
block also may contain titles for an individual table or chart, different from a title that
applies to the entire report (main section). BusinessObjects supports different types of
blocks such as a simple table, crosstab, or chart. A block is one component within a
section.

Variables and Cells

A cell contains either fixed text, formulas, or report variables. Cells that contain fixed
text such as a title or a picture are referred to as constants; the contents of the cell never
change no matter which data you are viewing. Cells whose contents change may be either
a formula or a report variable. Report variables are pointers to the columns of data. When
a report author builds a query, the author selects objects from the universe. These objects
become variables in a report. There are three types of report variables that correspond
directly to how the universe designer defines an object:

A dimension object is denoted with a blue cube and is typically textual information by
which you sort and analyze numeric measures. In the wine reports shown thus far, color,
vintage, varietal, and country are all dimension variables.

A measure is a number that you want to analyze; it is denoted by a pink sphere or circle.
Average Rating and Average Price are measure variables.

A detail provides additional information about a particular dimension. You may want to
see the information in a list report but will not want to use it to analyze measures by.
Phone number and street address are typical detail variables.

In a spreadsheet, each cell contains the actual value. In a BusinessObjects document, the
cell contains either the constant value or a formula that tells BusinessObjects where to
find the data value. By viewing the report structure, you can see the true contents of each
cell, as shown in Figure To view the report structure, use the pull-down menu to select
View | Report Structure.

The first cell is a text cell whose contents do not change, referred to as a constant. All the
other cells show formulas that are used to retrieve the data values for each report variable.
For example, the following formula will retrieve the individual colors of Red, White, and
Rose:

=
is the name of the report variable for this particular dimension object. The column and
row headings (, ) as well as the measure variable that will display individual data values
() make up the block. All these components together make up the report.
Components of a Document
One of the hardest concepts with BusinessObjects is the document itself. A document is
not a simple report, but rather, a set of components that eventually present a report. A
document contains the following components:

One or more data providers that are typically SQL queries that extract information from
source databases.

A result set in which the results of the queries are stored as a microcube. You can view
the results through Data Manager.

One or more formatted reports. Each report may be a different type, such as a chart, table,
or crosstab. One report may have multiple report types.

Figure gives a conceptual overview of a document that is made up of two data sources: a
SQL query and a spreadsheet. The document contains three reports, two that are tabular
reports with a view to each result set and a third that displays a chart with data from both
result sets. In many documents, you may have only one query, one result set, and one
report. Alternatively, you may have one query, one result set, and multiple reports. Each
report tab may contain a view with the full data set but in a different block type such as
table, crosstab, or chart. Alternatively, each report may contain a limited number of
columns or rows of data as you remove variables and apply filters. The structure of the
BusinessObjects document allows you to explore information from multiple perspectives
without ever having to requery the database. Similarly, the microcube technology allows
you to seamlessly combine information from multiple data sources into one report, even
if you don’t have a central data warehouse.
Home
BusinessObjects is a powerful ad hoc reporting and analysis tool. The single greatest
component of BusinessObjects that will make your implementation succeed or fail is the
universe. A universe is a business representation of your data warehouse or transaction
database. It shields users from the underlying complexities of the database schema.
BusinessObjects often refers to this as the semantic or metadata layer. In all your
development efforts, you must stay focused on that purpose: business representation. If
your universe becomes a glorified entity relationship model, your project will fail. If your
universe includes every data element any user may possibly want from now to eternity,
your project will fail.

Universes can become unwieldy for end users. Poorly defined joins will result in
unnecessarily slow queries. The universe is the most important component to get right.

Keep It Simple
This universe was difficult for the administrator to maintain and was overwhelming even
to expert BusinessObjects users. The result? End users often created invalid queries and
blamed the data warehouse for bad data. Casual users would ask an IT expert to create
MS Access data marts that were easier for them to use, thus defeating the flexibility and
empowerment offered by an ad hoc query tool and causing unnecessary data
reconciliation.

To build a successful universe, keep it simple. The universe should be useful for a clearly
defined group of users and should not have much more than 200 objects in it. Bigger
universes are technically feasible but not user friendly. Having more universes to build
and maintain may result in slightly higher maintenance costs but will significantly
increase end-user productivity and satisfaction. As your target user group expands,
constantly ask yourself if the needs are distinct enough to justify a separate universe. If
some users need only a handful of additional objects, keep it in the same universe.
However, if they need many additional objects, create a separate universe.
Figure illustrates how different user groups will need access to different information.
Human resources is one group of users that needs access to salary details but does not
need product sales and order information. Therefore, a Salary universe will only have
information from this one fact table. Marketing people may need information on sales but
will rarely need information on the individual order numbers that customer service
representatives need; however, customer service representatives need both order detail
and summary sales. This is an example in which it may make sense to have one universe
that meets the needs of both user groups (marketing and customer service
representatives). A director of the marketing group is most likely a people manager and
may need salary and employee details; the director would use two universes, as including
three subject areas in one big universe would potentially be overwhelming for the
majority of users who don't need this information.

Universe Components
BusinessObjects administrators build universes using Designer. The key components of a
universe are classes, objects, tables, joins, and contexts. As shown in Figure, classes and
objects are the main items a business user sees when building a query. Objects become
individual columns in a report; classes never appear in a report.

Classes
Classes are a way of grouping individual objects. In Figure, these appear with a folder
icon. Sometimes these relate closely to the tables in a database but should be regrouped
into business topics. In the sample EFASHION universe, the class Product is a more
meaningful business term than Article and includes items from multiple tables
ARTICLE_LOOKUP and ARTICLE_COLOR_LOOKUP.

Objects
Objects refer to columns of data. There are different types of objects (as explained further
in Chapter 8) denoted with a square, sphere, or triangle icon in Figure 5-2. Objects can
include a significant amount of intelligence and may not relate directly to one column in
the database. For example, the object Sold At (Unit Price) includes a calculation of
revenue/quantity. However, to avoid divide by 0 errors, it also includes an if-then-else
statement to check for 0 quantities. This is one example of why Business Objects
universes are so powerful and a much better alternative to providing users with direct
access to tables; if-then-else statements in SQL are implemented differently for each
RDBMS and are not something most users would know how to write.

Tables, Joins, and Contexts

Report authors never directly see several core elements of a universe: tables, joins, and
contexts (see Figure). Universe designers use tables to map data from fields to objects in
the universe. Joins allow the use of more than one table in a report, and contexts resolve
which join path to take when more than one path is possible. All three of these
components are then combined to dynamically build SQL statements in BusinessObjects.
Tables
Tables are individual database tables that provide data. A table may be a physical table in
the RDBMS, or it may be a view or synonym. Further, Designer provides functionality to
create aliases that are treated like tables.

In a data warehouse or data mart environment, you will have two types of tables: 1) a fact
table that contains numeric information and 2) dimension tables that allow a user to
analyze the numeric data from different perspectives such as product, time, or geography.
The fact table can have millions of detailed rows of data or can be smaller, with summary
numbers. One fact table together with its associated dimension tables is referred to as a
star schema. There can be multiple fact tables and star schemas within a universe.

Dimension tables are also referred to as lookup tables or reference tables. The dimension
tables can be broken into more than one table; for example, detailed material IDs may
reside in a MATERIAL_ID table. The groupings and product hierarchy for the material
IDs may reside in a separate table such as PRODUCT_GROUPING. This type of
structure, referred to as a snowflake design, is used in some data warehouses that have
extremely large dimensions as well as certain ROLAP tools.

In a normalized OLTP, both the fact tables and the dimension tables may be spread
across many tables. For example, order information may exist in both an
ORDER_HEADER table and an ORDER_LINES table. Dimensions and hierarchies
often do not exist in the OLTP (note in Figure that there is no Time or Plant table, just the
individual facility that produced the product). Only the individual material IDs, customer
IDs, and so on, are stored with detailed records. BusinessObjects does not allow a
universe to point to two different databases, so having data that users want to analyze
together generally calls for a data warehouse or data mart. However, if this is not
immediately possible, BusinessObjects provides a workaround in the end-user tool. You,
as the designer, can create two separate universes: one that points to the OLTP and one
that points to the dimension database. Users then would have to build two queries;
however, as long as the detailed key information is named consistently between the
universes, the results will be nicely displayed in one table, without the user having to
manually stitch the two result sets together.

When you build a universe, you are not replicating any data from these tables. Instead,
you are basically creating pointers to tell BusinessObjects where to find the data; no data
is stored in the universe itself. This is a drastically different approach than a full MOLAP
tool such as Hyperion Essbase, Cognos Powerplay, or Microsoft Analysis Services. Data
gets replicated only when a BusinessObjects user launches a report and the RDBMS
sends results back to the report, populating a micro cube in a .rep file on either the WebI
middle tier or the Windows client.

Joins
Joins specify how tables, views, synonyms, or aliases relate to one another. Joins allow a
user to combine information from two or more tables. For example, in the following
diagram, there are joins between ORDERS_FACT and the dimension table PLANT as
well as between ORDERS_FACT and the dimension table PRODUCTS. There are no
joins to the SUPPLIERS table. Without this join, a user is not able to determine which
suppliers provide various products. There are many types of joins.

Contexts
Contexts group related joins. A context may group a set of joins together for each star
schema. Without contexts, BusinessObjects would generate SQL that contained a loop.
Loops generally result in incorrect queries with fewer rows returned than expected.
Earlier versions of BusinessObjects supported queries that contained only one context. As
contexts were generally confusing for end users, they were best avoided. BusinessObjects
now allows one query to generate multiple SQL statements, one for each context. This
allows users to query multiple star schemas to create powerful business reports. Two
examples follow.

Days Sales Inventory (DSIs) How many days worth of inventory do you have according
to the daily sales volume? As shown in Figure, this query would involve two contexts,
one with all the joins for the star schema with a SALES_FACT table and a second
context with all the joins related to INVENTORY_FACT.