Introduction to Database Systems

Lecture 3: SQL Joins

(including most of Ch. 6.1-6.2)

1
 SQL Joins
• Can use data from multiple tables:

SELECT pname, price

FROM Product, Company
WHERE manufacturer = cname AND
country = ‘Japan’ AND
price < 150;
• This is a selection and projection of the “join” of the
Product and Company relations.

2
 Interpreting Joins
• A JOIN B produces one row for every pair of rows
– one row from A and one row from B

Cname Country Pname Price Manufacturer

Canon Japan SingleTouch 149.99 Canon

GizmoWorks USA Gizmo 19.99 GizmoWorks

PowerGizmo 29.99 GizmoWorks

(‘Canon’, ‘Japan’, ‘SingleTouch’, 149.99, ‘Canon’)

3
 Interpreting Joins
• A JOIN B produces one row for every pair of rows
– one row from A and one row from B

Cname Country Pname Price Manufacturer

Canon Japan SingleTouch 149.99 Canon

GizmoWorks USA Gizmo 19.99 GizmoWorks

PowerGizmo 29.99 GizmoWorks

(‘Canon’, ‘Japan’, ‘Gizmo’, 19.99, ‘GizmoWorks’)

4
 Interpreting Joins
• A JOIN B produces one row for every pair of rows
– one row from A and one row from B

Cname Country Pname Price Manufacturer

Canon Japan SingleTouch 149.99 Canon

GizmoWorks USA Gizmo 19.99 GizmoWorks

PowerGizmo 29.99 GizmoWorks

(‘Canon’, ‘Japan’, ‘PowerGizmo’, 29.99, ‘GizmoWorks’)

5
 Interpreting Joins
• A JOIN B produces one row for every pair of rows
– one row from A and one row from B

Cname Country Pname Price Manufacturer

Canon Japan SingleTouch 149.99 Canon

GizmoWorks USA Gizmo 19.99 GizmoWorks

PowerGizmo 29.99 GizmoWorks

(‘GizmoWorks’, ‘USA’, ‘SingleTouch’, 149.99, ‘Canon’)

6
 Interpreting Joins
• A JOIN B produces one row for every pair of rows
– one row from A and one row from B

Cname Country Pname Price Manufacturer

Canon Japan SingleTouch 149.99 Canon

GizmoWorks USA Gizmo 19.99 GizmoWorks

PowerGizmo 29.99 GizmoWorks

(‘GizmoWorks’, ‘USA’, ‘Gizmo’, 19.99, ‘GizmoWorks’)

7
 Interpreting Joins
• A JOIN B produces one row for every pair of rows
– one row from A and one row from B

Cname Country Pname Price Manufacturer

Canon Japan SingleTouch 149.99 Canon

GizmoWorks USA Gizmo 19.99 GizmoWorks

PowerGizmo 29.99 GizmoWorks

(‘GizmoWorks’, ‘USA’, ‘PowerGizmo’, 29.99, ‘GizmoWorks’)

8
 Interpreting Joins
• A JOIN B produces one row for every pair of rows
– one row from A and one row from B
Cname Country Pname Price Manufacturer
Canon Japan SingleTouch 149.99 Canon

GizmoWorks USA JOIN Gizmo 19.99 GizmoWorks

PowerGizmo 29.99 GizmoWorks

• This join produces 6 different rows

– in general, # rows in join is (# rows in A) * (# rows in B)
– number of rows often much smaller after selection…
– DBMS will do everything in its power to not compute A
JOIN B
9
 Interpreting Joins (2)
• Can think of a join in terms of code:

for every row C in Company

{ for every row P in
Product {
if (P.manufacturer = C.cname and
C.country = ‘Japan’ and
P.price < 150.00)
output (C.cname, C.country,
P.pname, P.price, P.category,
} P.manufacturer);
}
 Types of Joins
• We usually think of the selection as part of the join
– e.g., manufacturer = cname and country = ‘Japan’ and …
– called the “join predicate”

• Join without a predicate is cross product / cross join

• Special names depending on predicate

– natural join if “=“ between pairs of columns with same name
– with well chosen col names, many joins become natural

• These are “inner” joins. We will discuss outer later…

11
 (Inner) Joins
SELECT a1, a2, …, an
FROM R1, R2, …, Rm
WHERE Cond

for t1 in R1: (Nested loop

for t2 in R2: semantics)
...
for tm in Rm:
if Cond(t1.a1, t1.a2, …):
output(t1.a1, t1.a2, …, tm.an)

12
 (Inner) joins
Company(cname, country)
Product(pname, price, category, manufacturer)
– manufacturer is foreign key

SELECT DISTINCT cname

FROMProduct, Company
WHERE country = ‘USA’ AND category = ‘gadget’
AND manufacturer = cname

13
 (Inner) joins
SELECT DISTINCT cname
FROMProduct, Company
WHERE country = ‘USA’ AND category = ‘gadget’
AND manufacturer = cname

Product Company
pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

Camera Photo Hitachi Canon Japan

OneClick Photo Hitachi Hitachi Japan

14
 (Inner) joins
SELECT DISTINCT cname
FROMProduct, Company
WHERE country = ‘USA’ AND category = ‘gadget’
AND manufacturer = cname
Product Company
pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

Camera Photo Hitachi Canon Japan

OneClick Photo Hitachi Hitachi Japan

pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

15
 (Inner) joins
SELECT DISTINCT cname
FROMProduct, Company
WHERE country = ‘USA’ AND category = ‘gadget’
AND manufacturer = cname

Product Company
pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

Camera Photo Hitachi Canon Japan

OneClick Photo Hitachi Hitachi Japan

Not output because country != ‘USA’

(also cname != manufacturer)
16
 (Inner) joins
SELECT DISTINCT cname
FROMProduct, Company
WHERE country = ‘USA’ AND category = ‘gadget’
AND manufacturer = cname

Product Company
pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

Camera Photo Hitachi Canon Japan

OneClick Photo Hitachi Hitachi Japan

Not output because country != ‘USA’

17
 (Inner) joins
SELECT DISTINCT cname
FROMProduct, Company
WHERE country = ‘USA’ AND category = ‘gadget’
AND manufacturer = cname

Product Company
pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

Camera Photo Hitachi Canon Japan

OneClick Photo Hitachi Hitachi Japan

Not output because category != ‘gadget’ (and …)

18
 (Inner) joins
SELECT DISTINCT cname
FROMProduct, Company
WHERE country = ‘USA’ AND category = ‘gadget’
AND manufacturer = cname

Product Company
pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

Camera Photo Hitachi Canon Japan

OneClick Photo Hitachi Hitachi Japan

Not output because category != ‘gadget’

19
 (Inner) joins
SELECT DISTINCT cname
FROMProduct, Company
WHERE country = ‘USA’ AND category = ‘gadget’
AND manufacturer = cname

Product Company
pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

Camera Photo Hitachi Canon Japan

OneClick Photo Hitachi Hitachi Japan

Not output because category != ‘gadget’

20
 (Inner) joins
SELECT DISTINCT cname
FROMProduct, Company
WHERE country = ‘USA’ AND category = ‘gadget’
AND manufacturer = cname

Product Company
pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

Camera Photo Hitachi Canon Japan

OneClick Photo Hitachi Hitachi Japan

Not output because category != ‘gadget’ (with any Company)

21
 (Inner) joins
SELECT DISTINCT cname
FROMProduct, Company
WHERE country = ‘USA’ AND category = ‘gadget’
AND manufacturer = cname

Product Company
pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

Camera Photo Hitachi Canon Japan

OneClick Photo Hitachi Hitachi Japan

restrict to category = ‘gadget’

22
 (Inner) joins
SELECT DISTINCT cname
FROMProduct, Company
WHERE country = ‘USA’ AND category = ‘gadget’
AND manufacturer = cname

Product (where category = ‘gadget’) Company

pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

Canon Japan

Hitachi Japan

restrict to country = ‘USA’

23
 (Inner) joins
SELECT DISTINCT cname
FROMProduct, Company
WHERE country = ‘USA’ AND category = ‘gadget’
AND manufacturer = cname

Product (where category = ‘gadget’) Company (where country = ‘USA’)

pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

Now only one combination to consider

(Query optimizers do this too.)

24
 (Inner) joins
SELECT DISTINCT cname
FROMProduct, Company
WHERE country = ‘USA’ AND category = ‘gadget’
AND manufacturer = cname

Alternative syntax:

SELECT DISTINCT cname

FROMProduct JOIN Company ON
country = ‘USA’ AND category = ‘gadget’ AND
manufacturer = cname

Emphasizes that the predicate is part of the join.

25
 Self-Joins and Tuple Variables
• Ex: find companies that manufacture both products in
the ‘gadgets’ category and in the ‘photo’ category
• Just joining Company with Product is insufficient:
need to join Company with Product with Product
FROM Company, Product, Product
• When a relation occurs twice in the FROM clause we
call it a self-join; in that case every column name in
Product is ambiguous (why?)
– are you referring to the tuple in the 2nd or 3rd loop?

19
 Name Conflicts we used cname / pname
to avoid this problem

• When a name is ambiguous, qualify it:

WHERE Company.name = Product.name AND …

• For self-join, we need to distinguish tables:

FROM Product x, Product y, Company
• These new names are called “tuple variables”
– can think of as name for the variable of each loop
– can also write “Company AS C” etc.
– can make SQL query shorter: C.name vs. Company.name

27
 Self-joins
SELECT DISTINCT z.cname
FROMProduct x, Product y, Company z
WHERE z.country = ‘USA’
AND x.category = ‘gadget’
AND y.category = ‘photo’
AND x.manufacturer =
cname
AND y.manufacturer = cname;
Product Company
pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

SingleTouch photo Hitachi Hitachi Japan

MultiTouch photo GizmoWorks

28
 Self-joins
SELECT DISTINCT z.cname
FROMProduct x, Product y, Company z
WHERE z.country = ‘USA’
AND x.category = ‘gadget’
AND y.category = ‘photo’
AND x.manufacturer =
cname
AND y.manufacturer = cname;
Product Company
x pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

SingleTouch photo Hitachi Hitachi Japan

MultiTouch photo GizmoWorks

29
 Self-joins
SELECT DISTINCT z.cname
FROMProduct x, Product y, Company z
WHERE z.country = ‘USA’
AND x.category = ‘gadget’
AND y.category = ‘photo’
AND x.manufacturer =
cname
AND y.manufacturer = cname;
Product Company
x pname category manufacturer cname country

Gizmo gadget GizmoWorks GizmoWorks USA

y SingleTouch photo Hitachi Hitachi Japan

MultiTouch photo GizmoWorks

30
 Self-joins
SELECT DISTINCT z.cname
FROMProduct x, Product y, Company z
WHERE z.country = ‘USA’
AND x.category = ‘gadget’
AND y.category = ‘photo’
AND x.manufacturer =
cname
AND y.manufacturer = cname;
Product Company
x pname category manufacturer cname country z
Gizmo gadget GizmoWorks GizmoWorks USA
y SingleTouch photo Hitachi Hitachi Japan

MultiTouch photo GizmoWorks

restrict to country = ‘USA’

Not output because y.category != ‘photo’ 24
 Self-joins
SELECT DISTINCT z.cname
FROMProduct x, Product y, Company z
WHERE z.country = ‘USA’
AND x.category = ‘gadget’
AND y.category = ‘photo’
AND x.manufacturer =
cname
AND y.manufacturer = cname;
Product Company
x pname category manufacturer cname country z
Gizmo gadget GizmoWorks GizmoWorks USA

y SingleTouch photo Hitachi Hitachi Japan

MultiTouch photo GizmoWorks

Not output because y.manufacturer != cname

25
 Self-joins
SELECT DISTINCT z.cname
FROMProduct x, Product y, Company z
WHERE z.country = ‘USA’
AND x.category = ‘gadget’
AND y.category = ‘photo’
AND x.manufacturer =
cname
AND y.manufacturer = cname;
Product Company
x pname category manufacturer cname country z
Gizmo gadget GizmoWorks GizmoWorks USA

SingleTouch photo Hitachi Hitachi Japan

y MultiTouch photo GizmoWorks

x.pname x.category x.manufacturer y.pname y.category y.manufacturer z.cname z.country

Gizmo gadget GizmoWorks MultiTouch Photo GizmoWorks GizmoWorks 26USA

 Outer joins
Product(name, category)
Purchase(prodName, store) -- prodName is foreign key
SELECT Product.name, …, Purchase.store
FROMProduct, Purchase
WHERE Product.name = Purchase.prodName

Or equivalently:
SELECT Product.name, …, Purchase.store
FROM Product JOIN Purchase ON
Product.name = Purchase.prodName

But some Products may not be not listed. Why? 27

 Outer joins
Product(name, category)
Purchase(prodName, store) -- prodName is foreign key

If we want to include products that never sold,

then we need an “outer join”:

SELECT Product.name, …, Purchase.store

FROM Product LEFT OUTER JOIN
Purchase ON Product.name =
Purchase.prodName

28
 SELECT Product.name, Purchase.store
FROM Product JOIN Purchase ON
Product.name = Purchase.prodName
Product Purchase
Name Category ProdName Store

Gizmo gadget Gizmo Wiz

Camera Photo Camera Ritz

OneClick Photo Camera Wiz

36
 SELECT Product.name, Purchase.store
FROM Product JOIN Purchase ON
Product.name = Purchase.prodName
Product Purchase
Name Category ProdName Store

Gizmo gadget Gizmo Wiz

Camera Photo Camera Ritz

OneClick Photo Camera Wiz

37
 SELECT Product.name, Purchase.store
FROM Product JOIN Purchase ON
Product.name = Purchase.prodName
Product Purchase
Name Category ProdName Store

Gizmo gadget Gizmo Wiz

Camera Photo Camera Ritz

OneClick Photo Camera Wiz

Name Store

Gizmo Wiz

38
 SELECT Product.name, Purchase.store
FROM Product JOIN Purchase ON
Product.name = Purchase.prodName
Product Purchase
Name Category ProdName Store

Gizmo gadget Gizmo Wiz

Camera Photo Camera Ritz

OneClick Photo Camera Wiz

Name Store

Gizmo Wiz

39
 SELECT Product.name, Purchase.store
FROM Product JOIN Purchase ON
Product.name = Purchase.prodName
Product Purchase
Name Category ProdName Store

Gizmo gadget Gizmo Wiz

Camera Photo Camera Ritz

OneClick Photo Camera Wiz

Name Store

Gizmo Wiz

40
 SELECT Product.name, Purchase.store
FROM Product JOIN Purchase ON
Product.name = Purchase.prodName
Product Purchase
Name Category ProdName Store

Gizmo gadget Gizmo Wiz

Camera Photo Camera Ritz

OneClick Photo Camera Wiz

Name Store

Gizmo Wiz

41
 SELECT Product.name, Purchase.store
FROM Product JOIN Purchase ON
Product.name = Purchase.prodName
Product Purchase
Name Category ProdName Store

Gizmo gadget Gizmo Wiz

Camera Photo Camera Ritz

OneClick Photo Camera Wiz

Name Store

Gizmo Wiz

Camera Ritz

42
 SELECT Product.name, Purchase.store
FROM Product JOIN Purchase ON
Product.name = Purchase.prodName
Product Purchase
Name Category ProdName Store

Gizmo gadget Gizmo Wiz

Camera Photo Camera Ritz

OneClick Photo Camera Wiz

Name Store

Gizmo Wiz

Camera Ritz

Camera Wiz

43
 SELECT Product.name, Purchase.store
FROM Product JOIN Purchase ON
Product.name = Purchase.prodName
Product Purchase
Name Category ProdName Store

Gizmo gadget Gizmo Wiz

Camera Photo Camera Ritz

OneClick Photo Camera Wiz

Name Store

Gizmo Wiz

Camera Ritz

Camera Wiz

44
 SELECT Product.name, Purchase.store
FROM Product LEFT OUTER JOIN Purchase ON
Product.name = Purchase.prodName
Product Purchase
Name Category ProdName Store

Gizmo gadget Gizmo Wiz

Camera Photo Camera Ritz

OneClick Photo Camera Wiz

Name Store

Gizmo Wiz

Camera Ritz

Camera Wiz

45
 SELECT Product.name, Purchase.store
FROM Product LEFT OUTER JOIN Purchase ON
Product.name = Purchase.prodName
Product Purchase
Name Category ProdName Store

Gizmo gadget Gizmo Wiz

Camera Photo Camera Ritz

OneClick Photo Camera Wiz

Name Store

Gizmo Wiz

Camera Ritz

Camera Wiz

OneClick NULL

46
 SELECT Product.name, Purchase.store
FROM Product RIGHT OUTER JOIN Purchase
ON Product.name = Purchase.prodName
Product Purchase
Name Category
ProdName Store
Gizmo gadget
Gizmo Wiz
Camera Photo
Camera Ritz
OneClick Photo
Camera Wiz

Name Store Phone Foo

Gizmo Wiz

Camera Ritz

Camera Wiz

NULL Foo

47
 SELECT Product.name, Purchase.store
FROM Product FULL OUTER JOIN Purchase
ON Product.name = Purchase.prodName
Product Purchase
Name Category
ProdName Store
Gizmo gadget
Gizmo Wiz
Camera Photo
Camera Ritz
OneClick Photo
Camera Wiz

Name Store Phone Foo

Gizmo Wiz

Camera Ritz

Camera Wiz

OneClick NULL
l
NULL Foo
48
 Outer Joins
• Left outer join:
– Include the left tuple even if there’s no match
• Right outer join:
– Include the right tuple even if there’s no match
• Full outer join:
– Include both left and right tuples even if there’s no
match

• (Also something called a UNION JOIN, though it’s rarely used.)

• (Actually, all of these are used much more rarely than inner joins.)

49
 Join Examples
• See lec03-sql-joins.sql…

50