Database Development

Dr. Chen, Oracle Database System (Oracle) 1

Database Development

DDL

DML

DCL

JL_D.B.

ORACLE(SQL Components)

(Retrieve Data and Produce Information from Multiple Tables)


Chapter 9Joining Data from Multiple

Tables(p.284-296; p.312-330)

Jason C. H. Chen, Ph.D.Professor of MIS

School of BusinessGonzaga University

Spokane, WA 99258 [email protected]


Objectives

• Identify a Cartesian join• Create an equality join using the WHERE clause• Create an equality join using the JOIN keyword• Create a non-equality join using the WHERE

clause• Create a non-equality join using the JOIN…ON

approach


Objectives (continued)

• Create a self-join using the WHERE clause• Create a self-join using the JOIN keyword• Distinguish an inner join from an outer join• Create an outer join using the WHERE clause• Create an outer join using the OUTER

keyword• Use set operators to combine the results of

multiple queries


Creating additional Tables for chapter 9

• Run the following command for creating additional tables:SQL>start c:\oradata\chapter9\JLDB_Build_9.sql

• The following new four tables will be added to your database (see next slide for details):– Warehouses,– Publisher2, Publisher3 and– Employees


wh_id locationNUMBER(2) VARCHAR2(12)

ID Name Contact Phone

NUMBER(2) VARCHAR2(23) VARCHAR2(15) VARCHAR2(12)

ID Name Contact PhoneNUMBER(2) VARCHAR2(23) VARCHAR2(15) VARCHAR2(12)

EMPNO LNAME FNAME JOB HIREDATE DEPTNO MTHSAL MGR

NUMBER(4) VARCHAR2(20) VARCHAR2(15) VARCHAR2(19) DATE NUMBER(2) NUMBER(7,2) NUMBER(4)

pk

pk

pk

pk

warehouses

Publishe2

Publishe3

Employees

Publishe2 Publishe3

warehousesAdditional Database for chapter 9

Do we really need to create two additional Publisher tables (i.e., Publisher2 and Publisher3) ?


Purpose of Joins

• Joins are used to link tables and reconstruct data in a relational database

• Joins can be created through:– Conditions in a WHERE clause– Use of JOIN keywords in FROM clause


How many records will be in the Cartesian Join?

Figure 9-1 Results of a Cartesian Product


Cartesian Joins• Created by omitting joining condition in the

WHERE clause or through CROSS JOIN keywords in the FROM clause

• Results in every possible row combination (m * n)

• They are useful when– performing certain statistical procedures for

data analysis


Cartesian Join Example:Omitted Condition

Figure 9-3 Producing an unintentional Cartesian join

Q: Why unintentional Cartesian join is produced?

A: Because Oracle didn’t know what data the two tables had in common.

Q: How to solve the problem?

A: Use of Equality Joins (or inner/simple/natural).

-- chapter 9, Figure 9-3; p.288SELECT title, nameFROM books, publisher;

(70 rows produced)


Cartesian Join Example:CROSS (Cartesian) JOIN Keywords

Figure 9-4 Using the CROSS JOIN keywords

-- chapter 9, Figure 9-2; p.287SELECT isbn, title, location, ' ' CountFROM books, warehousesORDER BY location, title;

-- chapter 9, Figure 9-4; p.289SELECT isbn, title, location, ' ' CountFROM books CROSS JOIN warehousesORDER BY location, title;

(42 rows produced)

-- chapter 9, Figure 9-2b; p.287SELECT count (location)FROM books, warehousesORDER BY location, title;


Equality Joins

• Link rows through equivalent data that exists in both tables

• Created by:– Creating equivalency condition in the WHERE

clause– Using NATURAL JOIN, JOIN…USING, or

JOIN…ON keywords in the FROM clause


Joining Multiple Tables

• Join: combine data from multiple database tables using foreign key references

• SELECT field1, field2, ... FROM table1, table2 WHERE table1.joinfield = table2.joinfield AND search_condition(s);• If tables share field names, must prefix field in select

with table name (table1.field1, table2.field1)• Join condition: part of where clause indicating how

tables are related (table1.foreign_key = table2.primary key)

• Search conditions can be added to join condition using AND operator


Inner Join (cont.)

-- Exatra exampleSELECT s_id, s_last, s_first, student.f_id, f_lastFROM student, facultyWHERE student.f_id = faculty.f_id;

S_ID S_LAST S_FIRST F_ID F_LAST------ -------- -------- ---- --------1 Jones Tammy 1 Cox2 Perez Jorge 1 Cox3 Marsh John 1 Cox4 Smith Mike 2 Blanchard5 Johnson Lisa 4 Sheng6 Nguyen Ni 3 Williams6 rows selected.

Q: why “Brown” is not on the result?

Q: How many rows will be produced if “C-Join” is used?


Natural Join – Inner Join

• It can be used when the tables have a single commonly named and defined column.

-- Extra ExampleSELECT s_id, s_last, s_first, student.f_id, f_lastFROM student, facultyWHERE student.f_id = faculty.f_id;

-- Use NATURAL JOINSELECT s_id, s_last, s_first, f_id, f_lastFROM student NATURAL JOIN faculty;

S_ID S_LAST S_FIRST F_ID F_LAST------ -------- -------- ---- --------1 Jones Tammy 1 Cox2 Perez Jorge 1 Cox3 Marsh John 1 Cox4 Smith Mike 2 Blanchard5 Johnson Lisa 4 Sheng6 Nguyen Ni 3 Williams6 rows selected.


Equality Joins (Traditional Method): WHERE Clause Example

Figure 9-6 An equality join


Qualifying Column Names• Columns in

both tables must be qualified

Figure 9-7 A “column ambiguously defined” error

• Which table is “pubid” from?

• Is it from publisher or books?-- chapter 9, Figure 9-7(b); p.292

SELECT title, books.pubid, nameFROM books, publisherWHERE books.pubid = publisher.pubid;


WHERE Clause Supports Join and Other Conditions

Figure 9-8 Including search and join conditions in a WHERE clause

-- Use aliases -- chapter 9, Figure 9-9; p.294SELECT b.title, b.pubid, p.nameFROM books b, publisher pWHERE b.pubid = p.pubidAND (b.cost < 15 OR p.pubid = 1)ORDER BY title;


Using a Query Design Diagram• Helpful for creating complicated queries• Can use a formula to derive actual query from diagram

Customers• customer#

(j)• lastname (d)• firstname (d)

Orders• order# (j)• customer#

(j)

Orderitems• order# (j)• isbn (j)

Query: Display customers’ lastname, firstname and books’ title they purchased

Books• isbn (j)• title (d)

j: joins: searchd: display







(j)



You can derive your query from the diagram by following these steps:1. Place the display fields in the SELECT clause2. List all of the tables in the FROM clause3. Include the links in join conditions in the WHERE clause4. Include all of the search fields in the WHERE clause (if needed)

Figure: Join query design diagram







(j)




-- chapter 9, Figure 9-10; p.295SELECT c.lastname, c.firstname, b.titleFROM customers c, orders o, orderitems oi, books bWHERE c.customer# = o.customer#AND o.order# = oi.order#AND oi.isbn = b.isbnORDER BY lastname, firstname;


Joining More Than Two Tables

Figure 9-10 Joining four tables

• Joining four tables requires three join conditions

• Joining N tables requires ___ join conditions


Query: Display customers’ lastname, firstname and books’ title and only in ‘COMPUTER’ category


Customers• customer# (j)• lastname (d)• firstname (d)

Orders• order# (j)• customer# (j)


Books• isbn (j)• title (d)• _______


-- chapter 9, Figure 9-10; p.295SELECT c.lastname, c.firstname, b.titleFROM customers c, orders o, orderitems oi, books bWHERE c.customer# = o.customer#AND o.order# = oi.order#AND oi.isbn = b.isbn

ORDER BY lastname, firstname;AND category = ‘COMPUTER’

category (s)


Equality Joins: NATURAL JOIN

Figure 9-12 Using the NATURAL JOIN keywords


No Qualifiers with a NATURAL JOIN

Figure 9-13 Column qualifier error with a NATURAL JOIN

Natural Join keyword (e.g, pubid), we are not required to be specified when the two tables have it in common.

Therefore most developers avoid using a NATURAL JOIN because it can cause unexpected results.


Equality Joins: JOIN…USING

Figure 9-14 Performing a join with the JOIN … USING keywords


Break• Assignments (#2 & #5 on p. 329; see at the end of

the slides)• Figure out how to produce the output with ‘$’

displayed (below is a sample output for #5, p.329) – hint:

• a) column …• b) TO_CHAR (see p.365-367 and ‘learning to learn’)

TITLE PROFIT ------------------------------ -------- PAINLESS CHILD-REARING $37.45 HOW TO MANAGE THE MANAGER $16.55 PAINLESS CHILD-REARING $37.45

TO_CHAR( )PROFIT, _______


Query: list all author IDs with books in the ‘Children’ category

Query: list all author IDs with books in the ‘Family Life’ category

SELECT ba.authorid FROM books b JOIN bookauthor ba USING (isbn) WHERE category = 'FAMILY LIFE‘;

SELECT ba.authorid FROM books b JOIN bookauthor ba USING (isbn) WHERE category = 'CHILDREN';


Query: list all author IDs with books in the ‘Family Life’ or ‘Children’ category

-- chapter 9, Figure 9-28; p. 313(version 1)SELECT ba.authorid FROM books b JOIN bookauthor ba USING (isbn) WHERE category = ‘FAMILY LIFE’

Anything not appropriate in the output?

OR category = ‘CHILDREN’;

How to take care this type of problem in an easy way?

-- chapter 9, Figure 9-28; p. 313(version 2)SELECT _______ (ba.authorid) FROM books b JOIN bookauthor ba USING (isbn) WHERE category = 'FAMILY LIFE' OR category = 'CHILDREN';

What else we can achieve the same goal (and other more complicated situation) ?


Using Set Operators to Combine Query Results

• Performs set operations on outputs of two unrelated queries

• They all require that both queries– have the same number of display fields in the

SELECT statement, and that– each field in the first query has the same data

type as the corresponding column in the second query.

• <query 1> <SET Operator> <query 2>


A B

SET THEORY


C

SET THEORY

=A-B

=B-A

=A INTERSECT B=

All Regions =A UNION B(with C counted just once)

A-B B-A

A

BC


C

SET THEORY

A MINUS B=D

B MINUS A=E

A INTERSECT B=

A UNION B =

D E

A UNION ALL B = D+C+E+__C

CD+C+E

A

B


Oracle 11g/SQL Set Operator (Table 9-2)

• UNION (see Figure 9-28)– returns all rows from both queries, but ONLY displays

duplicate rows once• UNION ALL (see Figure 9-30)

– returns all (duplicate) rows from both queries, and displays ALL duplicate rows

• INTERSECT (see Figure 9-34)– returns all matching rows that are returned by both queries

• MINUS (see Figure 9-35)– returns all rows returned by the first query minus the

matching rows returned by the second query

i Use to select data from multiple tables not connected with foreign key relationships

i Used to combine the results of two or more SELECT statements


Query: list all author IDs with books in the ‘Family Life’ or ‘Children’ category

SELECT ba.authorid FROM books b JOIN bookauthor ba USING (isbn) WHERE category = 'FAMILY LIFE‘

SELECT ba.authorid FROM books b JOIN bookauthor ba USING (isbn) WHERE category = 'CHILDREN';

UNION

Figure 9-28 Producing an unduplicated combined list with the UNION set operator


Set Operators: UNION and UNION ALL Examples

What is the “difference” on the outputs?


Figure 9-34 Identifying overlapping values with the INTERSECT set operator

Query: list all customer numbers and those customers who have placed an order recently

SELECT customer# FROM customers

SELECT customer# FROM orders;

INTERSECT


Query: list all customer numbers but haven’t placed an order recently

Figure 9-35 Subtract result sets with the MINUS set operator

SELECT customer# FROM customers

SELECT customer# FROM orders;

MINUS


Query: A list of faculty members whose offices are in the BUS building. (Extra example)

FACULTYf_first (d)i f_last (d)i loc_id (j)

LOCATIONloc_id (j)i bldg_code (s)

SELECT f_first, f_last

FROM faculty, location

WHERE faculty.loc_id = location.loc_id

AND bldg_code = 'BUS';

MORE EXAMPLES on SET OPERATORSYou need to run the following command to make the example work:@ c:\oradata\NW_CW\northwoods.sql


Query: A list of faculty members who have taught a course in the BUS building.

FACULTYf_first (d)i f_last (d)i f_id (j)

LOCATIONloc_id (j)i bldg_code (s)

COURSE_SECTIONf_id (j)i loc_id (j)

SELECT DISTINCT f_first, f_lastFROM faculty, location, course_sectionWHERE faculty.f_id = course_section.f_idAND location.loc_id = course_section.loc_idAND bldg_code = 'BUS';


Query: A list of faculty members whose offices are in the BUS building or who have taught a course in the BUS building. (extra example)

SELECT f_first, f_lastFROM faculty, locationWHERE faculty.loc_id = location.loc_idAND bldg_code = 'BUS'

SELECT f_first, f_lastFROM faculty, location, course_sectionWHERE faculty.f_id = course_section.f_idAND location.loc_id = course_section.loc_idAND bldg_code = 'BUS';

UNION

Officein ‘BUS’

Taughtcoursesin ‘BUS’


Query: A list of faculty members whose offices are in the BUS building and who have taught a course in the BUS building. (extra example)



INTERSECT

Officein ‘BUS’



And NEXT ...

Query: A list of faculty members who have taught a course in the BUS building, but whose office are NOT located in the BUS



MINUS

Officein ‘BUS’



• Practice all the examples in the text.• A Script file is available on the Bb (file

name: ch9Queries.sql)• After completing all examples, do the HW.


Bonus Assignment

--****************************************************************-- Marketing department is analyzing books that don't sell. -- A list of ISBNs and Book Titles for all books is needed. -- Use a set operation (based on JL_Database to complete this task.--****************************************************************

-- a sample solutionISBN TITLE---------- ------------------------------0132149871 HOW TO GET FASTER PIZZA0299282519 THE WOK WAY TO COOK4981341710 BUILDING A CAR WITH TOOTHPICKS


Homework - Hands-On Assignments

Read and Practice all examples on Chapters 9• 1. Run the script files (in the folder \oradata\chapter9\):

JLDB_Build_9.sql• 2a. Read Oracle assignment and create a script file

Oracle_ch9_Lname_Fname.sql for questions (#2 & #5 ; p.329) on “Hands-on Assignments” (use TWO SQL queries, traditional one and with “JOIN” read instructions carefully)

• 2b. Be sure to use i) traditional method, ii) JOIN keyword, iii) draw Query Design Diagrams for each problem, and iv) use COLUMN statement to produce readable outputs – see next slide for details

• 3. Execute and test one problem at a time and make sure they are all running successfully.

• 4. When you done, spool the script files (see next slide for spooling instructions) and UPLOAD the WORD file including the Query Design Diagrams (see detail on the next slide) to Bb by the deadline.

Upload the WORD file (includes the Query Design Diagrams)to the Bb (under “Assignments & Projects”) by the deadline.


How to Spool your Script and Output Files and add Query Design Diagrams

After you tested the script file of Oracle_ch9_Lname_Fname.sql successfully, follow the instructions below to spool both script and output files:

Step 0. Run the following script file from SQL*Plus (since you have created JLDB tables)– Start c:\oradata\chapter9\JLDB_Build_9.sql

• 1. type the following on SQL>– Spool c:\oradata\Oracle_ch9_Spool_Lname_Fname.txt (make sure

your name is entered)• 2. open Oracle_ch9_Lname_Fname.sql that you already tested• 3. copy and paste all the SQL commands (including all comments) to the

SQL*PLUS • 4. type Spool Off on the SQL>The output should contain your personal information, all SQL commands and

their solution on the .txt file and saved in C: drive (oradata\folder). Be sure that COLUMN commands might be needed on the script file to align

the output.


How to Spool your Script and Output Files and add Query Design Diagrams (continued)

• 5. Next, you should use MS/Word to open the *.txt file and

include/DRAW “Query Design Diagram” figures for queries with multiple tables. You may use font of “Courier new” to align the output.

• 6. Save it as *.docx file.

Upload the WORD file (includes the Query Design Diagrams)to the Bb (under “Assignments & Projects”) by the deadline.


Summary• Data stored in multiple tables regarding a single

entity can be linked together through the use of joins

• A Cartesian join between two tables returns every possible combination of rows from the tables; the resulting number of rows is always m * n

• An equality join is created when the data joining the records from two different tables are an exact match

• A non-equality join establishes a relationship based upon anything other than an equal condition

• Self-joins are used when a table must be joined to itself to retrieve needed data


Summary (continued)

• Inner joins are categorized as being equality, non-equality, or self-joins

• An outer join is created when records need to be included in the results without having corresponding records in the join tables– The record is matched with a NULL record so it will be

included in the output• Set operators such as UNION, UNION ALL,

INTERSECT, and MINUS can be used to combine the results of multiple queries

Database Development

Documents