DB2 Guide

DB2

DB2

DB2 HO VER 0.3 1

DB2

Table of Contents

INTRODUCTION TO DB2...........................................................................................................................51.1 Data Base Management System.............................................................................................................5

1.1.1) Relational Data Base Management System...................................................................................51.2) D B 2.....................................................................................................................................................5

1.2.1) Structured Query Language (SQL)................................................................................................51.3) How does a COBOL program use DB2 services..................................................................................6

1.3.1) Host Variables................................................................................................................................61.3.2) SQL Communication Area.............................................................................................................61.3.3) SQL statements..............................................................................................................................6

1.4) Handling a request.................................................................................................................................6SYSTEM STRUCTURE................................................................................................................................8

2.1) Major Components................................................................................................................................82.1.1) System Services.............................................................................................................................82.1.2) Locking Services............................................................................................................................82.1.3) Database Services...........................................................................................................................8

2.1.3.1 ) Precompiler............................................................................................................................92.1.3.2) Bind.........................................................................................................................................92.1.3.3) Runtime Supervisor.................................................................................................................92.1.3.4) Data Manager..........................................................................................................................92.1.3.5) Buffer Manager.......................................................................................................................9

2.1.4) Distributed Data Facility................................................................................................................92.2) Detailed Control Flow.........................................................................................................................10

2.2.1) Syntax Checking..........................................................................................................................112.2.2 ) Optimization................................................................................................................................112.2.3) Package Generation......................................................................................................................112.2.4.) Authorization checking...............................................................................................................11

2.3) Packages Plans and Collections..........................................................................................................12DATA TYPES AND OPERATORS...........................................................................................................13

3.1) Data Types..........................................................................................................................................133.1.1)Numeric data.................................................................................................................................133.1.2) String Data...................................................................................................................................133.1.3) Date/Time.....................................................................................................................................13

3.2) Operators and Expressions..................................................................................................................143.3) Scalar Functions..................................................................................................................................14

DB2 DATA STRUCTURES........................................................................................................................164.1) Introduction.........................................................................................................................................164.2) Database..............................................................................................................................................164.3 ) Table Spaces.......................................................................................................................................16

4.3.1 ) Simple Table Space.....................................................................................................................164.3.2) Partitioned Table Space................................................................................................................174.3.3) Segmented Table Space...............................................................................................................17

4.4) DB2 Stored Tables..............................................................................................................................174.5) Index....................................................................................................................................................184.6) Index Spaces........................................................................................................................................194.7) Storage Groups....................................................................................................................................194.8) Views...................................................................................................................................................194.9) Alias....................................................................................................................................................204.10) The hierarchy of DB2 objects...........................................................................................................20

INTEGRITY.................................................................................................................................................215.1) Introduction.........................................................................................................................................215.2) Primary key.........................................................................................................................................21

5.2.1) Definition.....................................................................................................................................215.3) Entity Integrity....................................................................................................................................22

DB2 HO VER 0.3 2

DB2

5.4) Foreign Keys.......................................................................................................................................225.4.1) Definition.....................................................................................................................................22

5.5) Referential Integrity............................................................................................................................225.6) DELETE rules.....................................................................................................................................22

STRUCTURED QUERY LANGUAGE.....................................................................................................246.1) Introduction.........................................................................................................................................246.2) Types of SQL......................................................................................................................................24

DATA DEFINITION LANGUAGE...........................................................................................................257.1) Introduction.........................................................................................................................................257.2) CREATE TABLE...............................................................................................................................25

7.2.1 ) CREATE TABLE Using LIKE...................................................................................................267.3 ) ALTER TABLE.................................................................................................................................267.4 ) DROP TABLE...................................................................................................................................267.6 ) INDEXES...........................................................................................................................................27

7.6.1 ) CREATE INDEX........................................................................................................................277.6.2 ) DROP INDEX.............................................................................................................................27

DATA MANIPULATION LANGUAGE...................................................................................................288.1 ) Introduction........................................................................................................................................288.2 ) SELECT.............................................................................................................................................28

8.2.1) Simple Retrieval...........................................................................................................................288.2.2) Qualified retrieval........................................................................................................................298.2.3 ) Retrieval with Ordering..........................................................................................................298.2.4 ) Retrieval using BETWEEN and IN............................................................................................298.2.5 )Retrieval using LIKE...................................................................................................................308.2.6) Retrieval involving NULL...........................................................................................................30

8.3 ) JOIN Queries......................................................................................................................................318.3.1 ) Simple Equijoin...............................................................................................................................31

8.3.3 ) Greater than Join.........................................................................................................................318.3.4 )Join query with additional conditions..........................................................................................318.3.5 ) Joining a table to itself................................................................................................................31

8.4 ) SUBQUERIES...................................................................................................................................328.4.1 ) Simple Subquery.........................................................................................................................328.4.2) Correlated Subquery.....................................................................................................................328.4.3 ) Scalar Comparison Operator.......................................................................................................33

8.5 ) The Existential Qualifier....................................................................................................................338.6 ) Quantified Comparisons.....................................................................................................................338.7 ) Column Functions..............................................................................................................................348.8 ) GROUP BY........................................................................................................................................348.9 ) Use of HAVING................................................................................................................................348.10 ) UNION.............................................................................................................................................358.11 ) INSERT............................................................................................................................................358.12 ) UPDATE..........................................................................................................................................368.13) DELETE............................................................................................................................................36

THE CATALOG..........................................................................................................................................379.1 ) Introduction........................................................................................................................................379.2) SYSTABLES......................................................................................................................................379.3 ) SYSCOLUMNS.................................................................................................................................379.4 ) SYSINDEXES...................................................................................................................................379.5 ) QUERYING THE CATALOG..........................................................................................................379.6 ) ALIASES AND SYNONYMS..........................................................................................................38

VIEWS...........................................................................................................................................................3910.1 ) Introduction......................................................................................................................................3910.2) VIEW DEFINITION.........................................................................................................................3910.3) RETRIEVAL Operations..................................................................................................................3910.4) UPDATE Operations........................................................................................................................40

SECURITY AND AUTHORIZATION......................................................................................................41

DB2 HO VER 0.3 3

DB2

11.1 ) Introduction......................................................................................................................................4111.3 ) GRANT AND REVOKE.................................................................................................................4111.4 ) The GRANT OPTION.....................................................................................................................41

EMBEDDED SQL........................................................................................................................................4212.1 ) Introduction......................................................................................................................................4212.2 ) Preliminaries....................................................................................................................................4212.3 ) Operations Involving CURSORS....................................................................................................43

12.3.1 )DECLARE.................................................................................................................................4412.3.2 )OPEN.........................................................................................................................................4412.3.3)FECTCH.....................................................................................................................................4412.3.4)CLOSE........................................................................................................................................4412.3.5)WITH HOLD clause...................................................................................................................45

TRANSACTION PROCESSING................................................................................................................4613.1) Introduction.......................................................................................................................................4613.2) COMMIT and ROLLBACK.............................................................................................................4613.3) Three CONCURRENCY PROBLEMS............................................................................................47

13.3.1)The LOST UPDATE problem....................................................................................................4713.3.2)The UNCOMMITED DEPENDENCY problem........................................................................4713.3.3)The INCONSISTANT ANALYSIS problem.............................................................................48

13.4) Solution.............................................................................................................................................4813.5) Locking Facilities..............................................................................................................................49

13.5.1) LOCK TABLE...........................................................................................................................4913.5.2) The ISOLATION parameter......................................................................................................49

13.5.2.1)CURSOR STABILITY........................................................................................................4913.5.2.2) REPETABLE READ..........................................................................................................49

13.5.3)LOCK SIZE................................................................................................................................5013.5.4) The ACQUIRE/RELEASE parameters......................................................................................50

13.6) DEADLOCK.....................................................................................................................................50DYANAMIC SQL........................................................................................................................................51

14.1)Introduction........................................................................................................................................5114.2) PREPARE/EXECUTE......................................................................................................................5114.3) EXECUTE IMMEDIATE.................................................................................................................52

ADMINISTRATION FACILITIES............................................................................................................5315.1) Introduction.......................................................................................................................................5315.2) DB2 INTERACTIVE INTERFACE.................................................................................................53

15.2.1) SPUFI.........................................................................................................................................5415.2.2) DCLGEN....................................................................................................................................5415.2.3)Program Preparation....................................................................................................................5415.2.4)Precompile...................................................................................................................................5415.2.5) Bind / Rebind / Free...................................................................................................................5415.2.5) RUN...........................................................................................................................................5515.2.6) DB2 Commands.........................................................................................................................5515.2.7)Utilities........................................................................................................................................5515.2.8)Catalog Visibility........................................................................................................................5515.2.9) DB2I Defaults............................................................................................................................5515.2.10) EXIT.........................................................................................................................................55

15.3) EXPLAIN..........................................................................................................................................5515.4) Database utilities...............................................................................................................................56

DB2 HO VER 0.3 4

DB2

INTRODUCTION TO DB2

1.1 Data Base Management System

A Data Base, which consists of some collection of persistent data that is used by the application

systems of some given enterprise. These are designed and managed to store large quantities of information.

A Data Base Management System (DBMS) is a software package that manages the data stored in

databases. Data base approach will help to maintain data independence, avoid redundancy, inconsistency

and also provide better security. Thus DBMS provide an environment that is both convenient and efficient

to use in retrieving and storing data base information.

1.1.1) Relational Data Base Management System

A Relational Data Base System presents all information in form of tables. Tables consist of rows

and columns. This concept is in close correspondence with mathematical concept of relations and hence the

name.

1.2) D B 2

DB2 an abbreviation for DataBase 2, is a Relational DataBase Management System(RDBMS) that

runs on IBM mainframe computers under MVS operating system. Prior to DB2 there was a hierarchic Data

Base Management System (DBMS) for MVS called IMS (Information Management System). DB2 is one

of the most successful DBMS for MVS.

1.2.1) Structured Query Language (SQL)

This is the language by which request to DB2 are made.DB2 supports SQL, which provides

facilities for application programming as well as for data base management. SQL is both ANSI and ISO

standard. In fact it is the language used to access all relational databases.

The statements, which make up SQL are commonly categorized in to Data Definition Language

(DDL), Data Manipulation Language (DML), and Control statements. DBA uses DDL for defining various

objects. Application programmers uses DML to work with DB2. Control statements are used to handle

authorization to various resources.

DB2 HO VER 0.3 5

DB2

1.3) How does a COBOL program use DB2 services

When DB2 is used through COBOL, SQL statements are to be included in the program. Following

are the three things that are to be included when using DB2 in COBOL programs Host variables, SQL

communication Area, and SQL statements.

1.3.1) Host Variables

These are used for retrieving and putting data into DB2 tables. Host variables have to be defined

in the program, which is used to receive the data DB2 returns and from which DB2 can get the data to

update tables. They are called host variables because they reside in the storage owned by host program.

1.3.2) SQL Communication Area

DB2 provides the feedback about the success of each operation by Storing information in SQLCA.

So in COBOL program SQLCA have to be included.

1.3.3) SQL statements

SQL statements actually does the data base work required by the program. SQL

statements can be embedded into a programming language. These SQL statements are not understood by C,

COBOL etc. So before compiling the program, a preliminary process (Precompiler) must pull out the SQL

statements.

1.4) Handling a request

DB2 HO VER 0.3 6

DB2

When a request to update the database comes to DB2 from the user. The application manager

passes the userid to DB2. DB2 will check for the authorization of the user. It maintains a catalog to see

whether the user is authorized. If not authorized DB2 will return a code indicating it has refused the request

due to unauthorized user id. There are multiple concurrent processing handled by DB2. To avoid a possible

confusion DB2 puts a lock on data. DB2 maintains a log which keeps record about everything that is done

to the data. DB2 always writes to the log before writing to the data after the successful updation it will

mark a COMMIT to the log. Since the update is completed DB2 releases the lock and the operation

completed successfully

DB2 HO VER 0.3 7

DB2

SYSTEM STRUCTURE

2.1) Major Components

There are four major components, each of which divides into a number of subcomponents. They

are

The system service component

The locking services

The database services

The distributed data facility component

2.1.1) System Services

The system services supports the system operations. This handles DB2 system related tasks

including controlling connections to other MVS subsystems, handling system startup, shutdown and

operator communications, managing system log, gathering system-wide statistics, performance auditing

and accounting informations etc.

The system logs are data sets used to record the information for recovering the user and system

data in the event of system failure. The logs are of two types; active log and archive log.

2.1.2) Locking Services

Locking services are provided by an IMS Resource Lock Manager (IRLM), which is a general-

purpose locking manager. This is used to control the concurrent access to DB2 data.

2.1.3) Database Services

This is used to retrieval and updation of database i.e. implements the functions of SQL. This

includes five subcomponents. They are:

Precomipler

Bind

Runtime Supervisor

Data Manager

Buffer Manager

DB2 HO VER 0.3 8

DB2

2.1.3.1 ) Precompiler

This is a preprocessor for host programming languages. Precompiler analyses the host language

source module, stripping out the SQL statements and replace them by host language CALL statements. The

precompiler produces two output files. The first is intermediate source program, Which contains the

COBOL statements. The original SQL statements in the program will be commented out so that COBOL

compiler ignores them. Second file produced by precompiler called Data Base Request Module (DBRM).

2.1.3.2) Bind

This is used to bind a DBRM to produce a package and also it binds together a list of packages to

produce an application plan. The package can be thought as a set of internal control structures, representing

the compiled form of original SQL statements in the DBRM.

Collections are a set of logically related packages, which does not have a physical existence. All

the packages used in a given application are assigned to the same collection. Host language compiler,

which compiles the modified source code to produce the Object module for the host language.

2.1.3.3) Runtime Supervisor

This is resident in the main storage during the execution and it overseas the execution of job.

When the SQL statements are executed, control first goes to runtime supervisor, which uses the control

informations in the plan to request appropriate operations on the part of Data manager.

2.1.3.4) Data Manager

This is a very sophisticated access method. It performs the normal access method functions like

search, retrieval etc. It also invokes the other system components to invoke locking ,I/O operations etc.

2.1.3.5) Buffer Manager

This is a component responsible for transferring data between external storage and memory. It

uses several methods to optimize its performance so that amount of I/O can be reduced.

2.1.4) Distributed Data Facility

This component called DDF, which provide Distributed database support. It has to work with

other subsystems to attain this facility.

DB2 HO VER 0.3 9

DB2

2.2) Detailed Control Flow

This section will explain how to prepare and execute DB2 embedded application program. After preparing

the program with embedded DB2 statements it has to be precompiled first. This is because the host

language compiler does not understand the embedded DB2 statements. The precompiler replaces all SQL

statements with CALL statements. The SQL statements will be commented out. Precompiler uses SQL

statements to build a DBRM, which is stored in a partitioned dataset. DBRM contains edited form of SQL

DB2 HO VER 0.3 10

DB2

statements together with some additional informations. So the output of precompilation is modified source

code and DBRM.

Modified source code is compiled and linkedited in normal way by using the host language compiler.

Bind is really an optimizing compiler. It will convert SQL statements into optimized internal form. Its input

is DBRM and output is a package. The major functions performed by bind are

Syntax checking

Optimization

Package generation

Authorization checking

2.2.1) Syntax Checking

Bind examines the SQL statements for syntax errors.

2.2.2 ) Optimization

Bind includes an optimizer to choose for each SQL statement an optimal access strategy for

implementing that statement. For example when the user wants to access some data the access path is

selected by the optimizer. Bind will generate a code that is tightly bound to optimizer’s choice of path.

2.2.3) Package Generation

This is the actual process of building the package.

2.2.4.) Authorization checking

Bind checks for authorization. It check for whether the user is allowed to perform the operations

requested on DBRM and also whether allowed to assign packages applicable to packages collection.

Bind also performs the plan bind. The input of this process is a list of packages. Output is an

application plan, which is stored in DB2 directory. This also done some authorization checking such as

whether the owner is authorized to execute all application packages.

At the time of execution both load module (from modified source code) and application plan (from

DBRM) to be used. The load module is loaded to memory and it stars execution. When it reaches the first

call to DB2 language interface module it passes the control to runtime supervisor. The runtime supervisor

retrieves the application plan from DB2 directory and loads it into memory. Then the control is passed to

DB2 HO VER 0.3 11

DB2

Data manager, which performs the necessary operations on the data and passes the result back to the

program.

2.3) Packages Plans and Collections

Package is a compiled form of DBRM. Plan is a list of packages. Plan may consist of more than

one packages. So if a given DBRM has to be recompiled, the appropriate package has to be bind not the

entire plan. If multiple plans involve the same DBRM, that can be compiled once and the corresponding

package can be referenced multiple times instead of binding the DBRM each time.

Collections is a logical grouping of packages. It is even possible to add a new package to a plan

without binding it. This is by using the concept of collections. When a plan is bound the input can be a

collections. This is equivalent to adding all packages in that collection. Including packages in the collection

can be done after plan is bound. Whatever packages in the collection while plan is executed is considered to

be a part of the plan.

DB2 HO VER 0.3 12

DB2

DATA TYPES AND OPERATORS

3.1) Data Types

DB2 supports the following scalar data types

3.1.1)Numeric data

INTEGER Full word binary integer, 31 bits and sign

SMALLINT Half word binary integer, 15 bits and sign

DECIMAL(p,q) Packaged decimal with p digit and sign. Decimal

point at q digit from right

FLOAT(p) Floating point number

3.1.2) String Data

CHARACTER(n) Fixed length string of ‘n’ 8bit characters. Maximum

value of n is 254

VARCHAR(n) Varying length string up to ‘n’ 8bit characters Max

value of ‘n’ is 4K/32K

GRAPHIC(n) Fixed length string of ‘n’ 16bit characters Max

value of ‘n’ is 127

VARGRAPHIC(n) Varying length string upto ‘n’ 16bit characters. Max

value of ‘n’ is 24K/16K

3.1.3) Date/Time

DATE Eight unsigned packed decimal occupying 4 bytes

YYYYMMDD

TIME 6 unsigned packed decimal occupying 3 bytes

HHMMSS

TIMESTAMP Sequence of 20 unsigned packed decimal digits

YYYYMMDDHHMMSSNNNNNN

DB2 HO VER 0.3 13

DB2

3.2) Operators and Expressions

DB2 provide a number of scalar operators that can be used in making scalar expressions.

Numeric Operators DB2 supports numeric operators such as +,-,/,*

These can be used with date,time,timestamp etc

Comparisons =,>,<.<>,>=,<=, Even two dates can be compared

Concatenation || This is used to concatenate two character or two

graphic strings

A scalar expression consists of scalar operands to produce another scalar value. These can appear

in SQL statements with SELECT, HAVING, WHERE etc. There are six types of scalar expressions. They

are Numeric, Character, Graphic, date, time,and timestamp.

E.g. Numeric Expression

STATUS

WEIGHT = WEIGHT*23

E.g. Character string

PNAME

INITIALS || LASTNAME

3.3) Scalar Functions

Scalar functions applied to a single column or expressions and operate on single value. The result

will be transformed version of column or expression. These can be executed in the

SELECT,WHERE,HAVING etc.

CHAR Converts DB2 time,date,timestamp,decimal to character

DATE Converts a value to a DB2 date. It can be a positive integer,

timestamp etc.

DAY Returns day portion of DB2 timestamp

DAYS Converts DB2 timestamp into number of days since jan1

0001

DECIMAL Converts numeric value to a decimal

DIGITS Converts a number to a string of digits. Truncate - sign

FLOAT Converts numeric to a floating point value

HEX Converts a value to hexadecimal

HOUR Returns hour portion of time, timestamp etc.

DB2 HO VER 0.3 14

DB2

INTEGER Converts a number to an integer. Truncate decimal point

LENGTH Returns length of any column which may be null

NULLIF Returns null if two expressions are equal. If not return first

expression

MICROSECOND Returns microsecond portion of timestamp

MINUTE Returns minute portion of a time stamp

MONTH Returns the month portion of a time stamp

SECOND Returns the second portion of timestamp

STRIP Removes leading and trailing blanks from a scalar expression

SUBSTR Returns a specified portion of a character string

TIME Converts a value to a DB2 time

VARGRAPHIC Converts a character string to a graphic string

YEAR Returns year portion of date, timestamp etc.

VALUE For a null column returns a value instead of null

DB2 HO VER 0.3 15

DB2

DB2 DATA STRUCTURES

4.1) Introduction

In MVS environment, the relational database manager itself is considered as a subsystem. For

each subsystem there are some system objects. DB2 data structures are treated to as DB2 objects. Each

DB2 object is used to support the structure of the data being stored. This section will detail about various

DB2 objects.

4.2) Database

A database is a collection of logically related objects such as tables together with their associated

index and the spaces which contain these tables, and index. It thus consists of a set of tablespaces together

with a set of index spaces. For a stored tables all its associated index must be wholly contained in a single

database.

Objects are grouped together in a database for operational reasons. Operator can make the

database available or unavailable for processing. Database is logical and the data can be contained in single

disk or a set of disk.

4.3 ) Table Spaces

A table space can be thought as a logical (Not in physically adjacent areas) address space use to

hold tables. It is at the tablespace level that storage is actually allocated. The maximum size of tablespace is

64GB and it is divided into pages, which are written to or read from a DASD at the time of an operation.

There is no limit for number of tablespace in a database or number of databases. Table Space can be

divided into three types

Simple

Partitioned

Segmented

4.3.1 ) Simple Table Space

A simple table space can contain one or more stored tables but one is the best option. Since more

than one logically related tables are present, records for different tables can be clustered together to

improve the access times. But in the case of sequential search the system will be slowed down as it has to

DB2 HO VER 0.3 16

DB2

search not only the stored records of the required table but also of the other tables that happen to be mixed

up with first one. So one table per table space is the most satisfactory arrangement for simple table space.

4.3.2) Partitioned Table Space

This is used to store very big tables with large number of rows. In such tables it is operationally

difficult to handle the entire table as one unit. A partitioned table space contains one stored table,

partitioned. Individual partitions are independent so that they can be recovered independent of each other.

4.3.3) Segmented Table Space

This also contains more than one stored tables. But here it will not support cross table clustering

i.e. do not allow the records of different tables to be interleaved on a single page as in simple table space.

Table space is divided in to segments. No segment is allowed to contain records from more than one table.

Sequential access to a particular table is more efficient. This can handle variable length records. So these

are more efficient as compared to simple table space.

4.4) DB2 Stored Tables

This is a stored representation of a base table. It consists of a set of stored records, one for each

data row in the base table. A stored record is not identical to corresponding row of base table. It consist of a

byte string which include

A stored record prefix containing control information

Up to N storage fields where ‘N’ is the number of columns of base table

A length prefix for the field which are of variable length

A null indicator prefix indicating nulls

An encoded value of the actual data

There is no gap between the fields. The varying fields contain only the actual value. All internally stored

records are addressed by RID or Record ID. RIDs are unique within the containing table space. From the

figure shown RID contain two parts: The page number and the byte offset from the bottom of the page to

identify the required slot.

DB2 HO VER 0.3 17

DB2

4.5) Index An index can be defined as a DB2 tool used to locate the row that contains a given value. When a

specific row is required DB2 reads the index finds the pointer to that row and then read the page contains

the row. So Indexing provides uniform and predictable performance for retrieval operations.

Index consist of three levels: a root page, intermediate pages, leaf pages. The leaf level contains an

entry for each distinct value for the indexed field. Intermediate level contains an entry for each leaf page.

Thus root page, intermediate page, leaf page provides a direct and fast access to indexed data. A stored

table can have any number of associated index and hence any number of logical ordering on it.

DB2 HO VER 0.3 18

DB2

4.6) Index Spaces

An index space is used to store an index. The correspondence between index and index space is

always one to one. An index space is created automatically when an index is created. Index spaces can be

recognized and recovered independently.

4.7) Storage Groups

A storage group is a named collection of DASD volumes. Each table space and an index space

normally have a storage group associated with it. Within each storage group data is stored by using VSAM

Linear Data Set (LDS).

4.8) Views

This is a virtual table that consisting of SQL SELECT statements to access data from one or more

tables. A view never stores the data. When user wants to access the view, SQL statement that defines the

view will be executed and the user gets the result. So these can be called as a users table whereas the data

reside in the actual tables on which the view is based. This provides a simplified coding and more security

as the data, which are not required, will be omitted from the view.

DB2 HO VER 0.3 19

DB2

4.9) Alias

A locally defined name for table or a view, in the same local DB2 system or in a remote DB2

system. Alias gives DB2 location independence because an alias can be created for a table at a remote site,

thereby freeing the user from specifying the site that contains the data. Alias can be used as a global one

because it can be used by anyone.

4.10) The hierarchy of DB2 objects.

The following figure demonstrates the hierarchy of various DB2 objects.

DB2 HO VER 0.3 20

DB2

INTEGRITY

5.1) Introduction

The term integrity is used in database contexts to refer to accuracy, validity or correctness of data

in the database. This is very important and is handled by system than by user. Integrity constraints to be

specified during database definition. This will be stored in catalog and will be used by the system during

various operations. DB2 includes two general integrity rules. These are brought into picture using two keys

called primary key and foreign key.

5.2) Primary key

Primary key is a unique identifier for the table. Even that can be composite i.e. more columns can

be combined together to form a unique key. But in a table there can be more than one unique identifier. In

that case table has multiple candidate keys out of which only one will be the primary key and the others

called alternate keys. Primary keys are very important as it provide the basic row level addressing

mechanism. Thus by using the primary key system can pinpoint some specific rows in a required table.

5.2.1) Definition

Primary key can be defined when the base table created through CREATE TABLE or can be

added to an existing table through ALTER TABLE. Every column participating in primary key must be

explicitly declared as NOT NULL. After creating the primary key appropriate unique index has to be

defined. This index called primary index enforces the uniqueness of primary key.

To an existing table primary key can be added using ALTER TABLE command. A suitable

unique index must already exist for ALTER TABLE to succeed, and also the column for primary key

should be declared as NOT NULL.

E.g. CREATE TABLE NTP

(EMPNO CHAR(5) NOT NULL,

EMPNAME CHAR(20) ,

PRIMARY KEY (EMPNO));

CREATE UNIQUE INDEX IND1 ON NTP(EMPNO);

It is also possible to remove the primary key from an existing table.

E.g. ALTER TABLE NTP DROP PRIMARY KEY;

DB2 HO VER 0.3 21

DB2

5.3) Entity Integrity

Entity integrity rule specifies that Primary key should be unique and it should not accept a NULL

value. This is because these values serve to identify the entire row. In the case of composite primary key

each individual value of the primary key should be wholly non null.

5.4) Foreign Keys

A foreign key is a column in one table that matches values to the primary key in another table. The

foreign key can also be null. Each value of the foreign key must match with a corresponding value of

primary key. The table, which contains the foreign key, is called referencing table and the table which

contain the corresponding primary key is called referenced or target table.

5.4.1) Definition

The foreign key can be defined when the base table is created using CREATE TABLE. It can also

be added to or deleted from an existing table using ALTER TABLE. It is not necessary to create an index

on foreign key.

E.g. CREATE TABLE EMP

(EMP# CHAR(5) NOT NULL,

EMPNAME CHAR(20) ,

PRIMARY KEY (EMP#)

FOREIGN KEY FK1 (EMPNO) REFERENCES NTP);

ALTER TABLE EMP DROP FOREIGN KEY FK1;

5.5) Referential Integrity

Referential constraint limits a set of foreign key values to a set of primary key values. So

referential integrity enforces referential constraints. So the database must not contain an unmatched foreign

key value.

5.6) DELETE rules.

When a row contains the primary key is deleted necessary changes to be occurred for the matching

foreign keys. There are three options for this: CASCADE, SET NULL, RESTRICT

CASCADE when the target row is deleted from target table the row which contain the foreign key of

referencing table also get deleted.

DB2 HO VER 0.3 22

DB2

SETNULL When the target row is deleted the foreign key will be set to null.

RESTRICT The delete is restricted.

This rule to be specified when defining the foreign key. Default is RESTRICT.

E.g. CREATE TABLE EMP

(EMP# CHAR(5) NOT NULL,

EMPNAME CHAR(20) ,

PRIMARY KEY (EMP#)

FOREIGN KEY FK1 (EMPNO) REFERENCES NTP ON DELETE CASCADE);

DB2 HO VER 0.3 23

DB2

STRUCTURED QUERY LANGUAGE

6.1) Introduction

SQL is used by DB2 for operating on database. In this programmer has to specify what data he

want to retrieve and not how to do it. In SQL a single request can be formulated in a number of different

but functionally equivalent ways. For example a nested query can always be converted to an equivalent

join. SQL is a nonprocedural language. An SQL manipulation statement can operate on a table and the

result can be obtained on another table. One retrieval statement can retrieve multiple rows or one update

can update can change multiple rows. This feature is called relational closure, which makes relational

database much easier.

6.2) Types of SQL

SQL can be categorized based on functionality. On the basis of this, it is divided into three

The Data Control Language(DCL) provide the control statements that govern data security

The Data Definition Language(DDL) for creating and maintaining the data structure

The Data Manipulation Language(DML) for accessing and modifying the data

Another way to classify SQL is by execution type. There are two types

The production where the SQL is planned and executed. Here all the requirements are well known

The ad-hoc where its undefined until an immediate need is identified

Another classification is based on the existence.

Embedded SQL where the SQL statements are embedded in program

Stand-alone where only SQL statements are used. No program required

SQL can also be categorized based on Dynamism

Dynamic SQL is that which is changeable at runtime. This can be either typed from the terminal at

runtime or can be constructed by application programs algorithms

Static SQL is unchangeable at runtime. This will be embedded in application programs.

DB2 HO VER 0.3 24

DB2

DATA DEFINITION LANGUAGE

7.1) Introduction

Data definition statements can be divided into two broad classes: Logical and Physical. The end

user will be normal using the logical statements. The DDL statements are CREATE, ALTER, DROP. The

following table summarizes these operations on various DB2 objects

DB2 Objects CREATE ALTER DROP

Storage Group Y Y Y

Database Y N Y

Tablespace Y Y Y

Table Y Y Y

Index Y Y Y

View Y N Y

In this section logical statements are covered in detail. The principal logical DDL statements are

CREATE TABLE, ALTER TABLE, DROPTABLE,CREATE VIEW, DROP VIEW, CREATE

INDEX,DROP INDEX.

7.2) CREATE TABLE

This creates the base table. A base table consists of rows and columns to define physical

characteristics of data to be stored. The format is

CREATE TABLE <table name>

(column definition [, column definition]...

[, Primary key definition]

[,alternate key definition]

[,foreign-key definition ])

[,other parameters];

The table name which is user defined gives a unique identity to the table. The column definition can take

the form column data-type [NOT NULL/NOT NULL WITH DEFAULT/UNIQUE]

The square brackets are used to show that they are optional. System will take the default values where

applicable.

E.g. CREATE TABLE S

( S# CHAR(5) NOT NULL,

DB2 HO VER 0.3 25

DB2

SNAME CHAR(20) NOTNULL WITH DEFAULT,

STATUS SMALLINT NOT NULL WITH DEFAULT,

CITY CHAR(15) ,

PRIMARY KEY (S#));

Eg2. CREATE TABLE P

( P# CHAR(6) NOT NULL,

PNAME CHAR(20) NOTNULL WITH DEFAULT,

COLOR CHAR(6) NOTNULL WITH DEFAULT,

WEIGHT SMALLINT NOT NULL WITH DEFAULT,

CITY CHAR(15) ,

PRIMARY KEY (P#));

Eg3. CREATE TABLE SP

( S# CHAR(5) NOT NULL,

P# CHAR(6) NOT NULL,

QTY SMALLINT);

7.2.1 ) CREATE TABLE Using LIKE

A new table can be created with the same shape as some existing table using the option LIKE. But

the new table created will not inherit Primary key, Alternate key, Foreign key definitions.

E.g. CREATE TABLE SCOPY LIKE S;

This is same as CREATE TABLE SCOPY

( S# CHAR(5) NOT NULL,

SNAME CHAR(20) NOTNULL WITH DEFAULT,

STATUS SMALLINT NOT NULL WITH DEFAULT,

CITY CHAR(15));

7.3 ) ALTER TABLE

An existing table can be altered by using ALTER TABLE. For example a new column can be

added to an existing table.

E.g. ALTER TABLE NTP ADD EMPLEVEL SMALLINT;

This statement adds a column to the extreme right in the table NTP. This should not be declared as

NOTNULL. But ALTER cannot change the width or data type of an existing column.

7.4 ) DROP TABLE

DB2 HO VER 0.3 26

DB2

An existing table can be deleted using this option. The specified table is deleted from the system.

All indexes, views etc defined on this table will be dropped.

E.g. DROP TABLE NTP;

7.6 ) INDEXES

Indexes can be created and dropped using DDL statements. Any number of indexes can be created

on a base table. The two DDL statements are CREATE INDEX and DROP INDEX.

7.6.1 ) CREATE INDEX

The format for create is as follows

CREATE [UNIQUE] INDEX <index name>

ON <table name> (column[order],column[order);

The order refers to the order of the column, which has two values ASC,DESC represents ascending and

descending order. Default is ASC. Once the index is created it is maintained by the data manager.

E.g. CREATE UNIQUE INDEX IND1 ON EMP (EMP# DESC);

7.6.2 ) DROP INDEX

This is used to delete an index and the corresponding entry will be deleted from the catalog.

E.g. DROP INDEX IND1;

DB2 HO VER 0.3 27

DB2

DATA MANIPULATION LANGUAGE

8.1 ) Introduction

There are four SQL statements for manipulating the data that has been loaded into the tables. They

are SELECT,INSERT,UPDATE,and DELETE. SELECT retrieves the data,UPDATE changes the existing

values, DELETE removes rows from the table, INSERT adds new rows into the table. This section gives a

detailed description of all the four DML statements.

8.2 ) SELECT

SELECT is used to retrieve the data from the base table. The format is

SELECT [ALL/DISTINCT] <column names>

FROM <table name>

[WHERE <conditional expression>]

[GROUP BY <column names> ]

[HAVING <conditional expression>]

[ORDER BY <column names>];

The result of the above query will be another table derived from the base table.

8.2.1) Simple Retrieval

E.g. SELECT EMP#, EMPNAME, EMPLEVEL FROM EMP;

This query selects all the rows of table EMP. An alternate way of selecting the entire table is using *

E.g. SELECT * FROM EMP;

To retrieve some specific columns that column names can only me mentioned in SELECT

E.g. SELECT EMP# FROM EMP;

To eliminate duplicate rows DISTINCT can be used in SELECT statement

E.g. SELECT DISTINCT EMPNAME FROM EMP;

To retrieve computed values, it can be given on SELECT statement.

E.g. SELECT EMP#, EMPLEVEL + 1 FROM EMP; The resultant table will be containing the

computed value.

DB2 HO VER 0.3 28

DB2

8.2.2) Qualified retrieval

A specific record can also be retrieved using some conditions. The conditional expression following

WHERE can be simple comparison or multiple comparisons.

E.g. SELECT * FROM EMP WHERE EMP# = ‘1001’ ;

E.g2. SELECT * FROM EMP WHERE EMP# = ‘1000’ AND EMPNAME = ‘UMA’ ;

8.2.3 ) Retrieval with Ordering

Resultant table obtained as a result of a query can be in an ordered form. For this ORDER BY has

to be used with SELECT.

E.g. SELECT EMP#,EMPNAME FROM EMP WHERE EMPLEVEL = ‘7’

ORDER BY EMPNAME DESC;

Here the required sequence is arranged. The ORDER BY can take either ASC or DESC. ASC is default.

E.g. SELECT EMP#,EMPNAME

FROM EMP

WHERE EMPLEVEL = ‘7’

ORDER BY 2 , DESC;

Here this will produce the same result as above . 2 refer to second column of the resultant table.

8.2.4 ) Retrieval using BETWEEN and IN

To retrieve the values between some specified ranges BETWEEN is used.

E.g. SELECT EMP#,EMPNAME

FROM EMP

WHERE EMP# BETWEEN ‘7000’ AND ‘8000’ ;

The same result can be produced by another query

E.g. SELECT EMP#,EMPNAME

FROM EMP

WHERE EMP# >= ‘7000’

AND EMP# <= ‘8000’ ;

NOT BETWEEN can also be used to exclude a specific range. IN is used in the similar fashion where the

individual comparisons ORed.

E.g. SELECT EMP#,EMPNAME

FROM EMP

WHERE EMP# IN ( ‘8000’, ‘8001’, ‘8002’);

This will produce the same result as

E.g. SELECT EMP#,EMPNAME

DB2 HO VER 0.3 29

DB2

FROM EMP

WHERE EMP# = ‘8000’

OR EMP# = ‘8001’

OR EMP# = ‘8002’ ;

NOT IN can also be used to exclude the specific values.

8.2.5 )Retrieval using LIKE

LIKE can be used to search for a specific character or a string or a character within a string. But

when using the LIKE

The ‘-‘ stands for a single character.

The % character stands for any sequence of N characters

All other character stands for themselves

E.g. SELECT EMP#,EMPNAME

FROM EMP

WHERE EMPNAME LIKE ‘ J%’ ;

This will retrieve all rows having EMPNAME starting with J.

E.g. SELECT EMP#,EMPNAME

FROM EMP

WHERE EMPNAME LIKE ‘%J%’ ;

This will retrieve all rows having J anywhere in the EMPNAME.

E.g. SELECT EMP#,EMPNAME

FROM EMP

WHERE EMPNAME LIKE ‘---J’ ;

This will retrieve the rows with EMPNAME having 4 characters with J as ending one. Not LIKE can also

be used as in previous cases.

8.2.6) Retrieval involving NULL

To retrieve the rows involving NULL values

E.g. SELECT S#

FROM S

WHERE CITY IS NULL;

IS NULL ore IS NOT NULL can be used. But it is illegal to use CITY = NULL.

DB2 HO VER 0.3 30

DB2

8.3 ) JOIN Queries

This is generally used to retrieve data from more than one table.

8.3.1 ) Simple Equijoin

The comparison operator will be ‘=’ in this case.

E.g. Consider two tables S and P.

SELECT S.*, P.*

FROM S, P

WHERE S.CITY = P.CITY;

Resultant table will include a Cartesian product of all the rows having the city same. The above query can

also be written as

Eg2. SELECT * FROM S, P WHERE S.CITY = P.CITY;

8.3.3 ) Greater than Join

This is the case where > is used instead of =

E.g. SELECT *

FROM S, P

WHERE S.CITY > P.CITY;

8.3.4 )Join query with additional conditions

In the WHERE statement more conditions can be added using AND or OR

E.g. SELECT *

FROM S, P

WHERE S.CITY = P.CITY

AND S.STATUS > 20;

More number of tables can also be used. When using it all the table names have to be included in the

FROM statement and Conditions have to be modified accordingly.

8.3.5 ) Joining a table to itself

This query involves joining a table to itself. So here two references of same table are used. Here

two range variables A,B are used. These variables ranges over that specified table only.

E.g. SELECT A.S#, B.S#

DB2 HO VER 0.3 31

DB2

FROM S A,S B

WHERE A.CITY = B. CITY;

8.4 ) SUBQUERIES

Subqueries are nested SELECT. SELECT FROM WHERE is nested within another such

expression.

8.4.1 ) Simple Subquery

e.g. SELECT S#

FROM S

WHERE CITY IN

(SELECT CITY FROM P WHERE P# = ‘P2’);

This will evaluate the nested subquery first. It will get the value of CITY.

E.g. If for P2 the CITY= PARIS

Then the above query will be same as

SELECT S#

FROM S

WHERE CITY IN (‘PARIS’);

The subquery, which is shown above, can be rewritten with a join also. Multiple levels of nesting can also

be included. In the above example in second WHERE another nesting can be added.

8.4.2) Correlated Subquery

This is a type of subquery in which the inner and outer select statements are executed

simultaneously.

E.g., SELECT SNAME FROM S

WHERE ‘P2’ IN

( SELECT P# FROM SP WHERE S# = S. S#);

Here inner nested query requires the variable S# from table S, whose value changes when the system

examines the table S. For example initial value S1 is selected. From the inner SELECT value of P#

corresponding to S1 are obtained. Then it evaluates the outer query. The corresponding SNAME will be

selected. Thus the processing for S1 will be completed. Then it selects the next value of S# from the table S

and the processing continues. A correlated subquery is one, which whose values depend on some variable

that receives its value from outer query. Thus both the queries are simultaneously executed. So such a

subquery has to be evaluated repeatedly instead of once.

DB2 HO VER 0.3 32

DB2

Even in the correlated subquery inner and outer queries can refer to the same table

E.g. SELECT A. P#

FROM SP A

WHERE A.P# IN

( SELECT B.P# FROM SPB WHERE B.S# = A.S#);

8.4.3 ) Scalar Comparison Operator

When using the subquery scalar comparison operators such as =,> etc can also be used.

E.g. SELECT S# FROM S WHERE CITY =

( SELECT CITY FROM S WHERE S# = ‘S1’);

8.5 ) The Existential Qualifier

In existentially qualified condition ‘EXISTS (SELECT....FROM....) ‘ , the expression evaluates

true only when the subquery produces some result. It can be made clear from an example

E.g. SELECT SNMAE

FROM S

WHERE EXISTS

(SELECT * FROM SP

WHERE S# = S.S# AND P# = ‘P2’);

Consider the first SNAME. Let it be SMITH with S#=S1. For SMITH to appear in resultant table

existence test to evaluate to true. It checks for rows with S#= S1 and P# = P2. If it find some rows then

the subquery evaluates true. So the SNAME correspond to that S# can appear in resultant table.

So the EXISTS in SQL evaluates to true if subquery argument returns a nonempty set.

NOT EXISTS can also be used in a similar fashion.

8.6 ) Quantified Comparisons

A quantified comparison operator consists of a scalar comparison operator like = , >, etc followed

by ANY or ALL.

E.g. SELECT PNAME

FROMPWHEREWEIGHT>ALL

( SELECT WEIGHT FROM P WHERE COLOR = ‘BLUE’);

Here the subquery returns the weight of all parts with color blue. The outer SELECT returns the rows with

values greater than all the values of previous result.

Eg2. SELECT S# FROM S WHERE CITY = ANY

DB2 HO VER 0.3 33

DB2

(SELECT CITY FROM P );

8.7 ) Column Functions

Column (Aggregate ) functions, compute from a group of rows, a single value. This provides

capability to aggregate data so as to perform statistical calculations across many rows with one SQL

statement. The column functions are

AVG Average of the values in the column

SUM Sum of the values in the column

COUNT Number of values in the column

MAX Largest value in the column

MIN Smallest value in the column

Eg1. SELECT COUNT(*) FROM S;

This gives the number of rows in the table S.

Eg2. SELECT COUNT (DISTINCT(S#)) FROM S;

This won’t count repetitions.

Eg3. SELECT MAX(WEIGHT) FROM P;

Eg4. SELECT MIN(WEIGHT) FROM P;

Eg5. SELECT SUM(QTY) FROM SP;

8.8 ) GROUP BY

Using this table can be grouped for taking some aggregate values.

E.g. SELECT P#, SUM(QTY)

FROM SP

GROUP BY P#;

This query will give each value of P# along with the sum of QTY for that specific value of P#.

Eg2. SELECT P#, SUM(QTY), MAX(QTY)

FROM SP

WHERE S# <> ‘S1’

GROUP BY P#;

8.9 ) Use of HAVING

When using GROUP BY, to specify condition HAVING is used. So in short HAVING is to

groups is what WHERE is to rows.

DB2 HO VER 0.3 34

DB2

E.g. SELECT P#

FROM SP

GROUP BY P#

HAVING COUNT(*) > 1;

8.10 ) UNION

The UNION operates on two sets of rows and combines them together into a single set. Two sets

of rows must be union compatible. i.e. the two original sets must be of the same shape.

E.g. SELECT P# FROM P WHERE WEIGHT > 16

UNION

SELECT P# FROM SP WHERE S# = ‘S2’ ;

The duplicate rows will always be eliminated. But if the user wants all the rows to be retrieved he has to

use UNION ALL

E.g. SELECT P# FROM P WHERE WEIGHT > 16

UNION ALL

SELECT P# FROM SP WHERE S# = ‘S2’ ;

Any number of SELECT can be UNIONed together. If OREDER BY clause is used it should come as a

part of final SELECT only. In simple equijoin it select the matching rows only. But if any user want both

matching and unmatching rows from any table (Outer Join) UNION ALL can be used.

E.g. SELECT S.*,SP.P#

FROM S,SP

WHERE S.S# = SP.S#

UNION ALL

SELECT S.* ,’ ‘

FROM S

WHERE NOT EXISTS

(SELECT * FROM SP WHERE SP.S# = S.S#);

8.11 ) INSERT

This is used to insert data in to the table created. The format is

INSERT INTO <tablename> [(col1)[col2,col3,..])]

VALUES (literal[literal....])

The values can also be obtained from executing some subqueries.

E.g. INSERT INTO P

VALUES (‘P1’ , ‘CINTHOL’ , ‘YELLOW’ , 10, ‘MADRAS’);

This is case where the column names can be omitted as such since one entire row is inserted.

DB2 HO VER 0.3 35

DB2

E.g. INSERT INTO P (P#, CITY, WEIGHT)

VALUES (‘P7’, ‘ATHENS’ , 24);

This is the case where the values of whole row are not inserted. So Selectively values can be inserted.

E.g. INSERT INTO P (P#)

SELECT P# FROM SP WHERE S#= ‘S1’;

Here we are not directly entering the values but they are selected from table SP.

8.12 ) UPDATE

The UPDATE statement has the general form

UPDATE <tablename>

SET <column = value>

[WHERE <condition>]

This is used to update for the values in the table

E.g. UPDATE P

SET COLOR = ‘YELLOW’,

WEIGHT = WEIGHT + 5,

CITY=NULL

WHERE P# = ‘P2’ ;

This query will update the values as given where value of P# is P2.

E.g. UPDATE SP

SET QTY = 0

WHERE ‘LONDON’ =

( SELECT CITY FROM S

WHERE S.S# = SP. S#) ;

This uses a subqueruy

8.13) DELETE

The DELETE is used to remove row or rows that satisfy the specific condition.

E.g. DELETE FROM S WHERE S# = ‘S5’ ;

This is a single row delete as the specific condition is given. Multiple rows can also be deleted

using a single query

E.g. DELETE FROM SP WHERE QTY > 300 ;

Subquery can also be used similar to previous cases.

DB2 HO VER 0.3 36

DB2

THE CATALOG

9.1 ) Introduction

DB2 catalog can be treated as a system database that contains information about various objects

such as base tables, views etc. Catalog in a DB2 system consists of various system tables. User can

interrogate catalog using standard facilities of SQL language. In DB2 system the catalog consist of 38

system tables. Some of these catalog tables are described below

9.2) SYSTABLES

This catalog table contains a row for every base table in the entire DB2 system. It include the

informations such as NAME of the table, CREATED BY give the user who created the table, CREATOR

which mentions the owner of the table and many other fields.

NAME CREATED BY CREATOR COLCOUNT OTHERINFOS

S JACOB JACOB 4 ------------

P JACOB JACOB 5 -------

9.3 ) SYSCOLUMNS

This catalog table contains a row for every column of each table of the system.

NAME TBNAME TBCREATOR COLTYPE ------------

S# S JACOB CHAR -------------

SNAME S JACOB CHAR ---------------

----- ----- -------- --------- --------------

9.4 ) SYSINDEXES

This catalog table contains a row for every index in the system.

NAME CREATED BY CREATOR TBNAME TBCREATOR ----------------

nameofindex JACOB JACOB S JACOB ----------

---- ----------- ---------- -- ----------- --------------

DB2 HO VER 0.3 37

DB2

9.5 ) QUERYING THE CATALOG

The catalog can be queried using SQL statements

E.g. SELECT TBNAME, TBCREATOR FROM SYSIBM.SYSCOLUMNS

WHERE NAME = ‘S#’ ;

These can also be used for specific case where in a user wanted to find the total number of tables

created by him. Then

E.g. SELECT COUNT(*) FROM SYSIBM.SYSTABLES

WHERE CREATED BY = ‘ JACOB’ ;

The catalog cannot be updated using INSERT UPDATE and DELETE operations.

9.6 ) ALIASES AND SYNONYMS

Alias is an alternate name for a base table. These will help to access the table very easily. For

example if user A create a table X and user B wants to access that table then the fully qualified name

should be given

E.g. SELECT * FROM A.X;

But if user B can create an alias and use it

E.g. CREATE ALIASE T1 FOR A.X;

SELECT * FROM T1;

Will give the same result as in the first case. Even a third user can also use the same alias by giving a fully

qualified name

E.g. SELECT * FROM B.T1;

These can be removed by using DROP statement.

E.g. DROP ALIASE T1;

A synonym is similar to an alias. But the basic difference is that it is private to the user who creates it. Also

synonym unlike alias cannot refer to remote tables.

E.g. CREATE SYNONYM S1 FOR A.X;

This can only be used by user who creates it. Also the table should not be remote. Removing the

SYNONYM can be done by

E.g. DROP SYNONYM S1;

DB2 HO VER 0.3 38

DB2

VIEWS

10.1 ) Introduction

A view is a virtual table that consists of SQL SELECT statements, which can access the data from

base tables. A view never stores the data. But when the view is accessed the SQL statements are executed

to access the data. So the view can access data from more than one table. These views provide with several

advantages. They allow same data to be seen by different users in different ways. The users perception is

simplified because by using a view he can select only those columns that are required. More security can be

given as all the columns that are not visible in the view. View also provides with logical data

independence.

10.2) VIEW DEFINITION

The general syntax is

CREATE VIEW < name> [ ( column[---] ) ]

AS <subquery> ;

E.g. CREATE VIEW VIEW1 (P#, PNAME , CITY)

AS SELECT P#, PNMAE, CITY FROM P WHERE COLOR = ‘RED’ ;

This will create a view with name A.VIEW1 where A is the CREATOR of the view. When the above

subquery is executed the subquery is not executed but simply stored in the catalog (SYSVIEWS) under the

specified name.

The view can be dropped by using DROP option.

E.g. DROP VIEW VIEW1;

Then for the specified view the definition will be removed from the catalog. But when a base table is

dropped all view defined on the base table will be automatically dropped. There is no ALTER view

statement.

10.3) RETRIEVAL Operations

Retrieval operations on a view will be converted into equivalent operations on a base table. But in

some case like when a user tries to create a view column which is derived from something other than a

simple column of the base table it can fail.

E.g. CRETE VIEW PQ (P#, TOTQTY)

AS SELECT P# , SUM(QTY) FROM SP GROUP BY P# ;

DB2 HO VER 0.3 39

DB2

SELECT AVG(TOTQTY) FROM PQ;

This if by simple merging process will give

SELECT AVG (SUM(QTY)) FROM SP GROUP BY P#;

This will fail as SQL will not allow aggegrate functions to be nested. So instead of this what DB2 does is

materializing the view. For example

E.g. CREATE TABLE TEMPPQ (P#, TOTQTY);

INSERT INTO TEMPPQ (P#, TOTQTY)

SELECT P#,SUM(QTY) FROM SP GROUP BY P# ;

SELECT AVG(TOTQTY) FROM TEMPPQ ;

DROP TABLE TEMPPQ ;

So in this case nesting of aggregate functions does not occur.

10.4) UPDATE Operations

Not all views are updatable. This is because a view can be selected from more than one base table.

Also there are chances to affect the various integrity relations. So in such case views are not updateable.

But if in some views we will be able to preserve the tables integrity relations then they are updateable.

DB2 HO VER 0.3 40

DB2

SECURITY AND AUTHORIZATION

11.1 ) Introduction

The security has to be given for protecting against unauthorized disclosure alteration etc in the

database. Each individual user is assigned an authorization id and all the tables that are created by the user

will be owned by this id. Authorization subsystem which allows users having certain privilege to

selectively grant those privilege to others. There are different types of privileges

Table privilege which allows the user to perform operations on tables

Plan privilege which are concerned with authorization to execute a given PLAN

Data base privileges which is concerned with operations such as CREATE table in a Data base

User privilege which authorize the use of storage objects namely storage groups, tablespace and buffer

pools

System privilege for the operations such as creation of new database.

11.3 ) GRANT AND REVOKE

For performing a particular operations user must hold the appropriate authority or privilege. These

two commands are associated with granting and revoking of privileges.

The syntax of the grant statements is

GRANT privileges [ON [type] objects ] to users;

E.g. GRANT SELECT ON TABLE S TO <userid>;

Similarly the revoke statements can be used to revoke privileges.

E.g. REVOKE SELECT ON TABLE S FROM <userid>;

11.4 ) The GRANT OPTION

If a privilege is granted to user A using GRANT OPTIONS, then user A can grant the same

privilege to user B.

E.g. GRANT SELECT ON TABLE S TO A WITH GRANT OPTION;

Then user A can grant the same privilege to user B

E.g. GRANT SELECT ON TABLE S TO B WITH GRANT OPTION.

Now if the grant options to User A is revoked by a revoke command it will have a cascading effect i.e. all

the associated privileges granted by A to various users will also be revoked.

DB2 HO VER 0.3 41

DB2

EMBEDDED SQL

12.1 ) Introduction

SQL statements can also be used in application program. So such SQL statements which are

included in the host language program are called Embedded SQL statements. The host languages that

currently supported by DB2 are PL/I, COBOL,FORTAN,C,SYSTEM/370 Assembler language.

12.2 ) Preliminaries.

Embedded SQL statements should be prefixed by EXEC SQL. And should be terminated by a

termination symbol(e.g. END EXEC).

An executable statements can appear wherever an executable host language can appear.

SQL statements can include references to host variables. Host variables should be prefixed by a colon

which helps to identify them from SQL column names. These host variables used should declared in

BEGIN DECLARE SECTION and terminated by END DECLARE SECTION.

Host variables can appear in SQL data manipulation language where literals are permitted.(E.g. INTO

clause in SELECT, WHERE and HAVING clause where values to be compared etc.)

Any table used in the program can be declared by means of an EXEC SQL DECLARE TABLE to

make the program more self-documenting.

SQLCODE AND SQLSTATE are two special host variables to get the feedback information after the

SQL query executed. A SQLCODE zero means that statement executed successfully. If a positive

value then statements executed with some exceptional conditions and a negative value means that the

statement did not execute successfully. The SQLSTATE values consists of two character ‘class code ‘

followed by three character ‘subclass code’ .This will also helps to know whether the SQL statements

are executed successfully. So to find the status every SQL statement in the program should be followed

by a test on SQLCODE or SQLSTATE.

SQLCA or SQL Communication Area can be included so that it contains the declarations for both

SQLCODE and SQLSTATE. So either use INCLUDE SQLCA or declare SQLCODE and

SQLSTATE explicitly

Embedded SQL SELECT statement requires INTO clause specifying the host variable so that values

retrieved from database can be assigned. These host variables should have a data type compatible with

the SQL data type of columns from which the values are retrieved. Host variables and database can

have same name.

E.g. EXEC SQL

SELECT STATUS,CITY

DB2 HO VER 0.3 42

DB2

INTO :STATUS, : CITY

FROM S

WHERE S# = ‘S5’

END-EXEC;

This will select the values of STATUS and CITY in to the host variables and SQLCODE will be

set to 0. If there is no row with S#= S5 then SQLCODE will be set to 100. If there is more than a row with

S# = S5 then the program will show an error with SQLCODE negative. If the values retrieved to be null,

the user has to include an indicator variable. For example if STATUS is NULL then

E.g. EXEC SQL

SELECT STATUS,CITY

INTO :STATUS INDICATOR :STATUSIND, : CITY

FROM S

WHERE S# = ‘S5’

END-EXEC;

If the value to be retrieved to be NULL the indicator value will be set to -1 and host variable

remains unchanged. Indicator variable should be declared as 15-bit signed binary integers.

E.g. EXEC SQL

INSERT INTO P(P#, PNAME, WEIGHT)

VALUES (:PNO, :PNAME, :WEIGHT)

END-EXEC;

Eg1. EXEC SQL

UPDATE S

SET STATUS = STATUS + : NEWSTAT

WHERE CITY = ‘LONDON’

END-EXEC;

12.3 ) Operations Involving CURSORS

In the case of singleton SELECT only one row has to be selected and the retrieved values will be

moved into host variable. But in some cases more number of rows have to be selected. DB2 application

programs uses cursor to navigate through a set of rows returned by an embedded SQL SELECT statement.

Cursor is declared, opened, rows are fetched from cursor one row at a time and cursor is closed.

Four distinct operations or cursors are

DB2 HO VER 0.3 43

DB2

12.3.1 )DECLARE

Declare defines the cursor, gives it a name unique to the program in which it is embedded and

assigns and SQL statement to the cursor name. The DECLARE statement does not execute the SQL

statement but merely defines it.

12.3.2 )OPEN

Open is an executable statement. It reads the SQL search fields, execute the SQL statement and

sometimes builds the resultant table. It does not assign values to host variables.

12.3.3)FECTCH

Returns the data from the resultant table one at a time to the specified host variables. If the

resultant table is not built at cursor open time, it is built one by one. When cursors are used for retrieving

multiple rows FETCH statement should be coded in a loop. When no more rows to be retrieved FETCH

statement should return an SQLCODE of +100.

12.3.4)CLOSE

This statement releases all statements used by cursor.

E.g. WORKING-STORAGE SECTION.

01 MORE ROWS PIC X VALUE ‘Y’.

EXEC SQL

DECLARE CURSOR C1 FOR

SELECT S# ,SNAME,STATUS

FROM S

WHERE CITY = ‘LONDON’

END-EXEC.

PROCEDURE DIVISION.

-

-

EXEC SQL

OPEN C1

DB2 HO VER 0.3 44

DB2

END-EXEC.

IF SQLCODE < 0 PERFORM ERR-PARA.

PERFORM PARA1 UNTIL MORE ROWS = ‘N’.

EXEC SQL

CLOSE C1

END-EXEC.

ERR-PARA.

--

--

PARA1.

EXEC SQL

FETCH C1 INTO :S#, :SNAME, :STATUS

END-EXEC.

IF SQLCODE = +100

MOVE ‘N’ TO MORE ROWS.

IF SQLCODE < 0 PERFORM ERR-PARA.

More number of cursors can be opened at any time. There is no limit to the number of cursors permitted per

application program. The cursor should be declared first before doing any processing on it. Update and

Delete can also be done. These are CURRENT forms of UPDATE and DELETE but it cannot be used

when UNION,GROUP BY,ORDRER BY,DISTINCT etc are used. Also this is not used in the case of

subqueries.

E.g. EXEC SQL

DELETE FROM S

WHERE CURRENT OF C1

END-EXEC.

12.3.5)WITH HOLD clause.

When a COMMIT is issued by program all open cursors will be closed unless WITH HOLD

option should be specified. So the current position of cursor can be saved.

E.g. EXEC SQL

DECLARE CURSOR C2 WITH HOLD FOR

SELECT S#,SNAME

FROM S

WHERE CITY = ‘LONDON’

END-EXEC.

DB2 HO VER 0.3 45

DB2

TRANSACTION PROCESSING

13.1) Introduction

A Transaction is a term used to specify the logical unit of work. This can involve any number of

updates in between. Thus it is a sequence of several operations which converts one consistent state of

database into another, without necessarily preserving the consistency in between. Generally explaining if

one transaction reaches its normal termination with all the steps in between working properly, then the

updation will be made permanent. Instead if some SQL ERROR or some system failure occurs, every

changes made will be erased. So it can be said that either full updation or no updation at all,but partial

updation cannot be made. These are taken care by system component called Transaction Manager.

Whenever an UPDATE is done it is not done on the database. If something goes wrong with

system, updation may be revoked either by program or by system. The data in such condition is called as

tentative until either of these happens

COMMIT

ROLLBACK

These two are not really database operations like SELECT and UPDATE. They are instructions to

Transaction Manager. A term Synchronization point will be used (Abbreviated as SYNCHPOINT) which

represents the level between two consecutive transactions.

13.2) COMMIT and ROLLBACK

Any updation made in the system is guarantied once it is committed. This could be by way of an

explicit COMMIT given by the program or it could be given by the system, when the task normally

terminates i.e. A successful end of a transaction is signaled as a synchpoint is established. By the way of

COMMIT statement all open cursors will be closed and all record locks will be released.

The SQL COMMIT statement takes the form

COMMIT [work];

ROLLBACK can also be given explicitly in the program or will be issued by the system when the

task abnormally terminates. Here all updations done after the previous SYNCHPOINT will not be saved to

database. Here again all open cursors are closed and record locks are released.

The syntax is

ROLLBACK[work];

DB2 HO VER 0.3 46

DB2

13.3) Three CONCURRENCY PROBLEMS

DB2 being a shared system allows numerous users to access the same database simultaneously.

Any such kind of system requires some kind of control mechanism to ensure that concurrent transaction do

not interfere with each others operation. There are essentially three ways in which things can go wrong.

They are

The lost update problem

The uncommitted dependency problem

The inconsistent analysis problem

13.3.1)The LOST UPDATE problem

Here Consider the two transactions A and B. From the figure given Transaction A retrieves row R

at time t1. Transaction B retrieves row R at t2. Now Transaction A updates row R at time t3. Again

Transaction B updates row R at t4. So transition A’s update is lost at time t4 because Transaction B

overwrites it. So this describes the lost update problem.

TRANSACTION A TIME `TRANSACTION B

------ T0 -------------

FECTCH R T1 -------------

--------- T2 FETCH R

UPDATE R T3 ------------

----------- T5 UPDATE R

13.3.2)The UNCOMMITED DEPENDENCY problem

This occurs when one transaction is allowed to retrieve a row that has been updated by another

transaction and has not yet been committed to database by other transaction.

TRANSACTION A TIME `TRANSACTION B

------ T0 -------------

T1 UPDATE R

FETCH R T2 ------------

T3 ROLLBACK

13.3.3)The INCONSISTANT ANALYSIS problem

DB2 HO VER 0.3 47

DB2

The figure shows two transactions A and B operating on an account ACC1.Transaction A is

summing an account balance and transaction B is transferring an amount 10 from ACC3 to ACC1. The

result produced by A is incorrect. So if A writes back the result into database it would leave the database in

an inconsistent state. Since A has seen an inconsistent state of database it has performed an inconsistent

analysis. But this cannot be an uncommitted dependency as B commits all updates before A sees ACC3.

13.4) Solution.

For solving the concurrency issues technique called Locking is used. The idea of locking is that

when transaction needs to update an object it puts a lock on that object. The effect of this is to lock the

other transactions out of the object thereby preventing any changes on it. These locks can be either shared

lock or exclusive lock.

DB2 HO VER 0.3 48

DB2

13.5) Locking Facilities.

Though DB2 provides implicit locking mechanism locking can also be done by explicit

declarations.

13.5.1) LOCK TABLE

The syntax is

E.g. LOCK TABLE SP IN SHARE MODE.

Eg1. LOCK TABLE P IN EXCLUSIVE MODE.

Where the table should be a base table and not a view. This command acquires an S or X lock on

the entire table as per the command given. Acquiring such lock may not be instantaneous since the program

will have to wait for a time, when all table level locks will be removed. Once this is done table level locks

will be made and if it is a shared one then the other transactions will be able to get shared locks on any

levels.

13.5.2) The ISOLATION parameter

The ISOLATION parameter on BIND command specifies, isolation level for the application plan

being bound. The two options are

REPETABLE READ

CURSOR STABILITY

13.5.2.1)CURSOR STABILITY

If a transaction obtains addressability to a particular record by setting a cursor to point to it and

acquires a lock on that record. This lock is in the shared mode and now without updating it, the cursor

removes the addressability i.e. the lock here will be in S mode and it does not change to X mode. Now this

lock can be released without waiting for the synchpoint in the case of CS being used. So CS provide more

concurrency.

13.5.2.2) REPETABLE READ

Cursor Stability though giving more concurrency is not advisable from the

theoretical point of view. So Repeatable Read is made as the default and that is more preferable. Repeatable

Read holds the record level S lock (Shareable Lock) until the next synchpoint similar to the X lock. The

problem here is that the record may be changed by another transaction without the notice of the change

being made, as explained in the lost update problem. But this can be avoided by the Isolation level

Repeatable Read. Here if a transaction obtains an addressability to record under CS and if there is a

DB2 HO VER 0.3 49

DB2

possibility that the transaction may update the record, i.e. the cursor is defined FOR UPDATE, then DB2

issues a Update Lock (ULOCK) for the transaction instead of the X lock. If a transaction holds a U lock

and another transaction requests for a U lock or X lock the request will be put in the wait state. If the first

transaction updates the record, the U lock will be changed to X lock else it will be released when the

addressability is removed. The advantage here is formation of dead locks may be avoided.

13.5.3)LOCK SIZE

For a given table space, the lock size can be specified as PAGE, TABLE, TABLESPACE or ANY

depending on the factor in the CREATE or ALTER TABLE SPACE operation. When TABLE is specified,

locking will be in the table level and when PAGE is specified, locking will be in the page level, when ANY

is specified DB2 itself decides which unit of locking would be optimum. Mostly DB2 acquires page level

lock initially and when the number of page level lock reaches a installation specified threshold value, then

DB2 escalates the level locking to a higher factor.

13.5.4) The ACQUIRE/RELEASE parameters

These parameters in the bind statement specifies when a table space level lock will have to be

acquired or released. For ACQUIRE possible specifications are USE And ALLOCATE, for RELEASE it

is COMMIT and DEALLOCATE. USE means that the Lock is acquired on the first use and ALLOCATE

means that the locks are allocated when the program begins execution. COMMIT means that the table

space level locks are released at each synch point, and in DEALLOCATE the locks will be held until the

program terminates.

13.6) DEADLOCK

It appears when two or more transactions are simultaneously in the wait state when one transaction

waits for the resource held by another while the other waits for the resource held by the first transaction.

When such a condition occurs, there is no other alternative except to detect it and break it. The breaking

part involves choosing one of the deadlock transaction as the victim and either rolling it back or asking it

role back. This releases the deadlock.

DB2 HO VER 0.3 50

DB2

DYANAMIC SQL

14.1)Introduction

Dynamic SQL consists of a set of embedded SOL facilities, that are provided specifically to allow

the construction of generalized on line application. There may be a necessity some times to construct the

required SQL statements dynamically and then to bind and create those constructive statements

dynamically. DB2I is a generalized on-line application, which can accept any valid SQL statements . It uses

the facility of Dynamic SQL to construct suitable SQL statements corresponding to its input, bind and

execute this statements and to return messages and results back to the terminal.

There are different types of Non SELECT Dynamic SQL statements. They are

PREPARE/ EXECUTE

EXECUTE IMMEDIATE

14.2) PREPARE/EXECUTE

This class of Dynamic SQL uses PREPARE and EXECUTE to issue SQL statements. Non SELECT

Dynamic SQL cannot issue a SELECT statement.

Eg1.

WORKING-STORAGE SECTION.

-

-

EXEC SQL INCLUDE SQLCA END-EXEC.

01 STRING-VAR.

02 STRING-VAR-LEN PIC S9(4) COMP.

02 STRING-VAR-TEXT PIC X(100).

PROCEDURE DIVISION.

MOVE 39 TO STRING-VAR-LEN.

MOVE “DELETE FROM TAB1 WHERE DEPTNO = ‘A00’” TO STRING-VAR-TEXT.

EXEC SQL

PREPARE STMT1 FROM :STRING-VAR

DB2 HO VER 0.3 51

DB2

END-EXEC.

EXEC SQL

EXECUTE STMT1

END-EXEC.

14.3) EXECUTE IMMEDIATE

EXECUTE IMMEDIATE which combines the functions of both PREPARE and EXECUTE

statements.

E.g. WORKING-STORAGE SECTION.

-

-

EXEC SQL INCLUDE SQLCA END-EXEC.

01 STRING-VAR.

02 STRING-VAR-LEN PIC S9(4) COMP.

02 STRING-VAR-TEXT PIC X(100).

PROCEDURE DIVISION.

MOVE 39 TO STRING-VAR-LEN.

MOVE “DELETE FROM TAB1 WHERE DEPTNO = ‘A00’” TO STRING-VAR-TEXT.

EXEC SQL EXECUTE IMMEDIATE : STRING-VAR END-EXEC.

DB2 HO VER 0.3 52

DB2

ADMINISTRATION FACILITIES

15.1) Introduction

This falls into two categories. They are facilities for database and facilities for system

administration. The database administration involves the design implementation and maintenance of

individual DB2 database and associated application. Two DB2 features that can assist these tasks are

DB2I and EXPLAIN facility. The system administration covers the installation, monitoring, tuning and

control of overall DB2 system itself.

15.2) DB2 INTERACTIVE INTERFACE

Almost all functions of DB2 are available online through the DB2 interactive interface. This

provides the ability to execute SQL statements interactively and to invoke pre-written application

programs. Apart from this they also have ability to issue operator commands, the ability to invoke database

utilities and the ability to prepare application programs for execution. Though DB21 is the normal mode of

operation, it is always possible to use batch. TSO supports both batch and online application. The online

application work under the control of ISPF. DB2I itself is a TSO online application, which provides the

ability to execute SQL statements interactively by means of a component call SPUFI (SQL Processor Using

File Input).

The following are the sequence of events for invoking DB2I

1. LOGON to TSO in the normal way

2. From ISPF Primary option menu, select DB21 option

Here we get the DB21 main menu with

SPUFI

DCLGEN

Program Preparation

Pre-compile

Bind / Rebind / Free

RUN

DB2 Commands

Utilities

Catalog visibility

DB2 HO VER 0.3 53

DB2

DB2I Defaults

Exit

15.2.1) SPUFI

This reports the interactive execution of SQL statements from TSO terminals. The idea here is that

the user can create a text file containing one or more SQL statements then execute the statement through

SPUFI and then use ISPF browse to browse the result.

15.2.2) DCLGEN

This allows the user to invoke the Declaration Generator program. This is the program that creates

embedded SQL, DECLARE TABLE statements and corresponding host language structure declarations

from table description in the catalog. The output from this is stored as a modules of a PDS which could be

copied by the include statement or to put up in the program itself.

15.2.3)Program Preparation

This allows the user to perform any of the following or all of the following functions.

Precompilation

Compilation or assembly

Linkage editor processing

Bind processing

Program execution (TSO application only)

15.2.4)Precompile

This is effective only for the precompile step of the program preparation .

15.2.5) Bind / Rebind / Free

BIND creates the application plan from one or more DBRM'S. The following parameters may be specified.

If the new plan is to replace the existing one

The AUTH ID that is to own the plan if that ID is different from the ID of the user entering the bind

command.

The isolation level RR & CS specifies the lock acquisition and release times for the plans.

DB2 HO VER 0.3 54

DB2

REBIND rebinds the existing plan. It is not a replace version of the bind command. Here the input for the

rebind command is the previous plan but not the DBRM.

15.2.5) RUN

RUN allows the user to execute a previously prepared application program.

15.2.6) DB2 Commands

This allows the user to enter system operator commands like START DATABASE, STOP

DATABASE etc.

15.2.7)Utilities

This allows the user to invoke various utilities

15.2.8)Catalog Visibility

Catalog Visibility menu allows certain queries to be formulated against DB2 catalog with out

direct use of SQL.

15.2.9) DB2I Defaults

This menu allows the user to set certain defaults for DB2I session

15.2.10) EXIT

This option allows the user to come out of DB2I session.

15.3) EXPLAIN

This allows the user to obtain information regarding the operators choice for the access strategy

specified for the SQL statement.

E.g. EXPLAIN PLAN FOR

SELECT SNO, NAME

DB2 HO VER 0.3 55

DB2

FROM TAB

WHERE CITY = PUNE

15.4) Database utilities

The DB2 database utilities are all online utilities and can be invoked online without stopping the

execution of DB2 system. Some of them are.

CHECK This utility checks a table space to find if there are any referential integrity violation. Records

if found to contain any unmatched foreign key value, the table space will be in the check pending state.

COPY This utility creates a full or incremental back-up copy of the table space. An incremental copy is

a copy of the pages that have changed since the previous change

LOAD This utility loads data from a sequential data set into one or more tables with one or more

partitions of a specified table space.

MERGE COPY This utility merges a full copy and one or more incremental copies to produce a upto

date full copy.

MODIFY This is used to perform maintenance operation recovery information in the DB2 catalog.

RECOVER This uses the most recent full copy, any subsequent incremental copies, and any

subsequent log data to recover one or more table space partitions or pages after a media failure

REPAIR This can be used to reset conditions like check pending.

REPORT This produces some reports that are useful in connection with managing database recovery

process

RUNSTATS This computes statistics on specified stored data and writes them to the system catalog.

STOSPACE This gathers information regarding the amount of space allocated to a given set of table

space and indexes and stores them in a DB2 catalog.

DB2 HO VER 0.3 56