Review

SQL: The Query Language

Part 1 • Relational Algebra (Operational Semantics)
• Given a query, how to mix and match the
relational algebra operators to answer it
R &G - Chapter 5
• Used for query optimization
• Relational Calculus (Declarative Semantics)
• Given a query, what do I want my answer set to
include?
• Algebra and safe calculus are simple and powerful
The important thing is not to models for query languages for relational model
stop questioning.
– Have same expressive power
Albert Einstein • SQL can express every query that is expressible in
relational algebra/calculus. (and more)

Query Optimization Relational Query Languages

• Two sublanguages:
Rel. Algebra
Query 1 – DDL – Data Definition Language
• Define and modify schema (at all 3 levels)
Rel. Algebra – DML – Data Manipulation Language
SQL Query Query 2 • Queries can be written intuitively.
.
. • DBMS is responsible for efficient evaluation.
.
– The key: precise semantics for relational queries.
Rel. Algebra
Query n – Optimizer can re-order operations, without affecting
query answer.
– Choices driven by cost model: how many disk
Pick the accesses; how much CPU?
cheapest one

The SQL Query Language Example Database

• The most widely used relational query language.

• Standardized Sailors Boats
sid sname rating age bid bname color
(although most systems add their own “special sauce”
-- including PostgreSQL) 1 Fred 7 22 101 Nina red
2 Jim 2 39 102 Pinta blue
• We will study SQL92 -- a basic subset 3 Nancy 8 27 103 Santa Maria red

Reserves
sid bid day
1 102 9/12
2 102 9/13
 Sailors
Review: SQL DDL The SQL DML sid sname rating age
sid sname rating age 1 Fred 7 22
CREATE TABLE Sailors ( 1 Fred 7 22
sid INTEGER, 2 Jim 2 39
sname CHAR(20), 2 Jim 2 39
3 Nancy 8 27
rating INTEGER, 3 Nancy 8 27
age REAL,
PRIMARY KEY sid); • Find all 18-year-old sailors:
bid bname color
CREATE TABLE Boats (
bid INTEGER, 101 Nina red
bname CHAR (20), 102 Pinta blue
SELECT * SELECT *
color CHAR(10)
PRIMARY KEY bid); 103 Santa Maria red FROM Sailors S FROM Sailors
WHERE S.age=18 WHERE age=18
CREATE TABLE Reserves (
sid INTEGER,
sid bid day
bid INTEGER,
day DATE, • To find just names and ratings, replace the first line:
1 102 9/12
PRIMARY KEY (sid, bid, date),
FOREIGN KEY sid REFERENCES Sailors, 2 102 9/13
FOREIGN KEY bid REFERENCES Boats); SELECT S.sname, S.rating

Querying Multiple Relations

Basic SQL Query
SELECT S.sname DISTINCT: optional keyword indicating target-list : A list of attributes
answer should not contain duplicates. of tables in relation-list
FROM Sailors S, Reserves R In SQL, default is that duplicates
WHERE S.sid=R.sid AND R.bid=102 are not eliminated! (Result is called
a “multiset”)
SELECT [DISTINCT] target-list
FROM relation-list
WHERE qualification
Sailors Reserves
sid sname rating age sid bid day
qualification : Comparisons
1 Fred 7 22 1 102 9/12 relation-list : A list of relation
combined using AND, OR and
2 Jim 2 39 2 102 9/13 NOT. Comparisons are Attr op names, possibly with a range-
3 Nancy 8 27 const or Attr1 op Attr2, where op is variable after each name
one of =,<,>,≠, etc.

Query Semantics SELECT [DISTINCT] target-list Find sailors who’ve reserved at least one
relation-list
FROM
boat
WHERE qualification

1. FROM : compute cross product of tables. SELECT S.sid

2. WHERE : Check conditions, discard tuples that fail. FROM Sailors S, Reserves R
3. SELECT : Delete unwanted fields.
WHERE S.sid=R.sid
4. DISTINCT (optional) : eliminate duplicate rows.

• Would adding DISTINCT to this query make a

Note: Probably the least efficient way to compute a query!
difference?
– Query optimizer will find more efficient ways to get the
same answer. • What is the effect of replacing S.sid by S.sname
in the SELECT clause?
– Would adding DISTINCT to this variant of the query
make a difference?
 About Range Variables Arithmetic Expressions
• Needed when ambiguity could arise.
– e.g., same table used multiple times in FROM
(“self-join”)
SELECT S.age, S.age-5 AS age1, 2*S.age AS age2
SELECT x.sname, x.age, y.sname, y.age FROM Sailors S
FROM Sailors x, Sailors y WHERE S.sname = ‘dustin’
WHERE x.age > y.age

Sailors SELECT S1.sname AS name1, S2.sname AS name2

sid sname rating age FROM Sailors S1, Sailors S2
1 Fred 7 22 WHERE 2*S1.rating = S2.rating - 1
2 Jim 2 39
3 Nancy 8 27

String Comparisons Intermission

SELECT S.sname
FROM Sailors S
WHERE S.sname LIKE ‘B_%B’ Why are Databases useful?
Here’s why
`_’ stands for any one character and `%’ stands for 0 or
more arbitrary characters.

Yes, every other language in the world uses Perl-like regular expressions.
In fact, PostgreSQL supports this with substring(), but this is not
standard or portable.

Find sid’s of sailors who’ve reserved a red or a green boat Find sid’s of sailors who’ve reserved a red and a green
boat

SELECT R.sid
FROM Boats B, Reserves R
WHERE R.bid=B.bid AND
(B.color=‘red’ OR SELECT R.sid
B.color=‘green’) FROM Boats B,Reserves R
... or: WHERE R.bid=B.bid AND
SELECT R.sid (B.color=‘red’ AND B.color=‘green’)
FROM Boats B, Reserves R
WHERE R.bid=B.bid AND
B.color=‘red’
UNION
SELECT R.sid
FROM Boats B, Reserves R
WHERE R.bid=B.bid AND B.color=‘green’
 Find sid’s of sailors who’ve reserved a red and a green Find sid’s of sailors who’ve reserved a red and a green
boat boat

• Could use a self-join:

SELECT S.sid
FROM Sailors S, Boats B, Reserves R
WHERE S.sid=R.sid
AND R.bid=B.bid SELECT R1.sid
FROM Boats B1, Reserves R1,
AND B.color=‘red’
Boats B2, Reserves R2
INTERSECT WHERE R1.sid=R2.sid
SELECT S.sid AND R1.bid=B1.bid
FROM Sailors S, Boats B, Reserves R AND R2.bid=B2.bid
WHERE S.sid=R.sid AND (B1.color=‘red’ AND B2.color=‘green’)
AND R.bid=B.bid
AND B.color=‘green’

Find sid’s of sailors who have not reserved a boat Nested Queries: IN

Names of sailors who’ve reserved boat #103:

SELECT S.sid SELECT S.sname
FROM Sailors S FROM Sailors S
WHERE S.sid IN
EXCEPT (SELECT R.sid
FROM Reserves R
SELECT S.sid WHERE R.bid=103)
FROM Sailors S, Reserves R
WHERE S.sid=R.sid

Nested Queries: NOT IN Nested Queries with Correlation

Names of sailors who’ve reserved boat #103:

Names of sailors who’ve not reserved boat #103:
SELECT S.sname
SELECT S.sname FROM Sailors S
FROM Sailors S WHERE EXISTS
WHERE S.sid NOT IN (SELECT *
(SELECT R.sid FROM Reserves R
FROM Reserves R WHERE R.bid=103 AND S.sid=R.sid)
WHERE R.bid=103)
• Subquery may need to be recomputed for each Sailors tuple.
– Think of subquery as a function call that runs a query!
• Also: NOT EXISTS.
 More on Set-Comparison Operators A Tough One
• we’ve seen: IN, EXISTS
Find sailors who’ve reserved all boats.
• can also have: NOT IN, NOT EXISTS
• other forms: op ANY, op ALL

SELECT S.sname
• Find sailors whose rating is greater than that of FROM Sailors S Sailors S such that ...
some sailor called Horatio: WHERE NOT EXISTS ( SELECT B.bid
there is no boat B without
FROM Boats B ...
SELECT *
WHERE NOT EXISTS ( SELECT R.bid
FROM Sailors S
a Reserves tuple showing S reserved B FROM Reserves R
WHERE S.rating > ANY
WHERE R.bid=B.bid
(SELECT S2.rating
AND R.sid=S.sid))
FROM Sailors S2
WHERE S2.sname=‘Horatio’)

Summary

• Relational model has well-defined query semantics

• SQL provides functionality close to basic relational model

(some differences in duplicate handling, null values, set
operators, …)
• Typically, many ways to write a query
– DBMS figures out a fast way to execute a query,
regardless of how it is written.