DB2 UDB FOR Z/OS

Overview

DB2 Application Programming

Using Embedded SQL

Program Structure

Overview

Why we need Embedded SQL ?.

Question

Where to Embedded SQL Statement ?.

Question

Host Languages

COBOL

PL/1
C Assembler FORTRAN

Supporting Embedded SQL

Use delimiters to identify SQL statements in the application

Use host variables to provide values on the WHERE clause or receive values from SELECT statements
Use appropriate techniques to process null data

Use communications from DB2 to determine success or failure of SQL statements

Embedded SQL Types

Overview

Embedded SQL Types

Static SQL

SQL Statements are presented in Program with same

function and same table

Dynamic SQL

SQL Statements are called at Run Time , Every time new

Function , New Table

Embedded SQL

Delimited all SQL statements

Declare a Communication Area

Describe Host Variables Code SQL statements to access DB2 data Handle Exceptional Conditions

SQL Delimiters

Used by PRECOMPILER to identify SQL statements to be translated.

Must indicate beginning and end of each embedded SQL statement.

The EXEC SQL must not be coded before column 12.

COBOL comments are allowed any where within the delimiters

Host Variables and Declaration

Overview

Host Variables

OPTIONAL, used to allow setting a value before statement is executed Host language label in SQL statement preceded by a colon Host variable must match column data type

Not usable for DB2 names (Tablename, columnname, . . .)

Host Variables Example

MOVE 2120 TO WS-EMPNO

EXEC SQL

SELECT EMPNO, LASTNAME,SEX

INTO :WS-EMPNO, :WS-LAST,:WS-SEX FROM EMPLOYEE

WHERE EMPNO= : WS-EMPNO

END-EXEC

Host Variables - Insert - COBOL Program

Host Variables - Update - COBOL Program

Host Variables - To Receive a Value

Receiving a Value - COBOL Program

Declaring Host Variables Language Definitions

Declaring COBOL Host Variables

All COBOL host variables must be explicitly declared in the DATA DIVISION.

A colon ( : ) must precede all host variables in an SQL statement The names of the host variables must be unique within the program.

Host variables, other than indicator variables, cannot be defined as arrays.

An OCCURS clause must only be specified when defining an indicator structure.

SQL Data Types Mapped to Typical COBOL Declarations

COBOL and SQL data types: Datetime

DATE TIME

PICX(10) PICX(08)

- Format specified by an option - Format specified by an option

TIMESTAMP PIC X (26) String representation of timestamp has the format: yyy-mm-dd-hh.mm.ss.nnnnnn

Host Structure for COBOL Program

MOVE 2120 TO WS-EMP.

Working-Storage Section. 01 WS-EMP-STRUCT.

EXEC SQL SELECT EMPNO, LAST, SEX INTO : WS-EMP-STRUCT FROM EMPLOYEE WHERE EMPNO = :WS-EMP

05 WS-EMPNO PIC S9(9) USAGE COMP. 05 WS-LAST PIC X (30). 05 WS-SEX PIC X(01).

END-EXEC.

DCLGEN (Generating Host Variable )

Example COBOL DCLGEN output:

Example COBOL DCLGEN output:

How to include Host variable Declaration and SQL Communication Area predefined Declaration in Program
Overview

INCLUDE
EXEC SQL

INCLUDE SQLCA | SQLCA |member-name

END-EXEC

Member-name

Member of the partitioned data set . The member can contain any COBOL variable declarations, source statements and any SQL statements other than an INCLUDE statement.

SQLCA (SQL Communication Area)

Overview

SQLCA

SQLCODE and SQLSTATE

SQLSTATE is set by the database manager after execution of each SQL statement. Application programs can check the success of the execution of SQL statements by testing SQLSTATE instead of SQLCODE.

SQLCA ( SQLCODE & SQLSTATE) Examples

What is SQLCA Warring and why we require in Embedded SQL program

Question

SQLCA Warnings

Example Program

Overview

SPUFI screen to request DCLGEN:

Embedded SQL Program to Insert data into Employee Table

Overview

DCLGEN-Input

DCLGEN- Output

DCLGEN- Output

DCLGEN - Output

EMBEDED SQL INSERT-COBOL Program

EMBEDED SQL INSERT-COBOL Program

EMBEDED SQL INSERT-COBOL Program

DB2 PRE-COMPILATION

BIND JCL

BIND JCL

Program Preparation

What This Unit Is About

How the Embedded SQL Program is going to be Processed by

PRECOMPILER
COMPILER BINDER PLAN and PACKAGE

DB2 Program Preparation

Four Major steps for Program Preparation

PRECOMPILE COMPILE

LINK-EDIT
BIND & RUN

What does DB2 Pre-compile does when you submit a Embedded SQL Program as Source ?.
Question

DB2 PRE-COMPILE Process

PRE-COMPILE Process
Embedded SQL Program has Application Code SQL Statements Before the program can be
compiled by the appropriate source code language compiler.

EXEC SQL statements

embedded in the program must be removed and replaced by

CALL Statements in COBOL

DB2 Pre-compile Process

PRECOMPILER prepares the source program for compilation by replacing EXEC SQL by a CALL and commenting out the EXEC SQL statement.

Verifies the SQL syntax.

PRECOMPILER process will include any files included with an

EXEC SQL INCLUDE

statement

Including files produced

by DCLGEN.

What is the Purpose of Database Request Module (DBRM) in DB2 Program Preparation ?.
Question

DBRM (Database Request Module )

COBOL Compile and LinkEdit Process

Overview

Compile and Link

After DB2 PRE-COMPILATION the modified Cobol program with CALL statements ( instead of EXEC SQL) is now ready to be compiled and link-edited. LINK-EDIT process will have to include the necessary modules for the call to work properly. Modified Embedded Cobol source file is

Compiled and linked

normally.

Produce an executable file

or load module.

Compile and Link

After DB2 PRECOMPILATION the modified Cobol program with CALL statements ( instead of EXEC SQL) is now ready to be compiled and link-edited. LINK-EDIT process will have to include the necessary modules for the call to work properly. Modified Embedded Cobol source file is

Compiled and linked

normally.

Produce an

executable file or load module .

Program Preparation : Bind Process

Overview

Bind Process

When the bind file or DBRM is bound to the database, a package is created in the database. package contains instructions about how to execute the SQL statements that were included in the source code.

Bind Process

Run

Run

Run Process

Accepts Input
Executes Load Module Retrieves Information from DB2 Package when program call SQL statements

Bind Plan and Package

Embedded SQL Program to Update data into Employee Table

Overview

EMBEDED SQL Update-COBOL Program

EMBEDED SQL Update-COBOL Program

EMBEDED SQL Update-COBOL Program

PRE-COMPILE

BIND JCL

OUTPUT

Embedded SQL Program to Delete data into Employee Table

EMBEDED SQL DELETE-COBOL Program

EMBEDED SQL DELETE-COBOL Program

EMBEDED SQL DELETE-COBOL Program

PRE-COMPILE

BIND JCL

OUTPUT

Embedded SQL Program to SELECT data FROM Employee Table

Overview

EMBEDED SQL SELECT-COBOL Program

EMBEDED SQL SELECT-COBOL Program

EMBEDED SQL SELECT-COBOL Program

PRE-COMPILE

BIND JCL

BINDJCL

BINDJCL

OUTPUT

OUTPUT

Special Registers
EXEC SQL SET :WS-TS = CURRENT TIMESTAMP END-EXEC. EXEC SQL SET : WS-USER = USER END-EXEC.

EXEC SQL SET CURRENT SQLID = : WS - SQLID END - EXEC.

Embedded SQL Program to handle NULL values (INDICATOR Variables )

Overview

Indicator Variables
Indicator variables are small integers that are used to :

Indicate whether the values of the associated host variables are NULL Verify that the value of a retrieved character string has not been truncated

Insert and update NULL values from host variables into columns.

Indicator Variables Example - COBOL Program

INDICATOR VARIABLES

A variable of SMALLINT data type associated with a host variable Must be declared. Is preceded by a colon and directly follows its associated host variables, when used in an SQL statement. EXEC SQL SELECT INTO FROM WHERE END EXEC.

HIREDATE :HV HIREDATE : IV HIREDATE EMPLOYEE EMPNO = :HV EMPNO

Note *

IV HIREDATE is the indicator variable associated with the host variable HV - HIREDATE

INDICATOR VARIABLES

DB2 returns one of the following values to an indicator variable:

if the vale returned to the associated host variable by DB2

is NOT NULL, not truncated

-1

is NULL

>0 is a truncated character string; the value in the indicator variable is the length of the character string before truncation.

Receiving NULL Values

MOVE ZEROES EXEC SQL

TO WS-LASTNAME-IND.

SELECT LASTNAME

INTO :WS-LAST :WS-LASTNAME-IND

FROM EMPLOYEE WHERE EMPNO = :WS-EMPNO END-EXEC. IF WS-LASTNAME-IND < ZEROES ..... END-IF.

Updating Columns with NULL

MOVE
MOVE

2120
SPACES

TO WS-EMPNO.
TO WS-PHONE-NO. TO WS-PHONE-NO-IND

MOVE - 1

EXEC SQL UPDATE EMPLOYEE SET PHONE = :WS-PHONE-NO :WS-PHONE-NO-IND WHERE EMPNO = :WS-EMPNO END-EXEC.

Inserting NULL values

MOVE -1 TO WS-PHONE-NO-IND.

EXEC SQL INSERT INTO EMPLOYEE (EMPNO, SEX, AGE,PHONE) VALUES ( :WS-EMPNO , :WS-SEX, :WS-AGE, :WS-PHONE-NO :WS-PHONE-NO-IND) END-EXEC.

Indicator Variables

Updating Valid Values

MOVE 2120 MOVE 616161 TO WS-EMPNO TO WS-PHONE-NO

MOVE ZEROES
..... UPDATE-PARA.

TO WS-PHONE-NO-IND.

PERFORM UPDATE-PARA.

EXEC SQL
UPDATE EMPLOYEE SET PHONE = :WS-PHONE-NO :WS-PHONE-NO-IND WHERE EMPNO = :WS-EMPNO

END-EXEC.

EXAMPLE Embedded SQL Program to handle NULL values (INDICATOR Variables )

Overview

EMBEDED SQL Handling NULL Values

EMBEDED SQL Handling NULL Values

EMBEDED SQL Handling NULL Values

EMBEDED SQL Handling NULL Values

PRE-COMPILE

BIND JCL

output

OUTPUT

SQL CODE
EXEC SQL SELECT EMPNO, LASTNAME, SEX INTO :WS-EMPNO, :WS-LAST, :WS-SEX

FROM EMPLOYEE WHERE EMPNO= :WS-EMPNO END-EXEC. EVALUATE SQLCODE WHEN 0 DISPLAY LAST NAME IS WS-LAST WHEN 100 DISPLAY WS-EMPNO DOES NOT EXIST END-EVALUATE

Processing Multiple Rows

Processing Multiple Rows

Structured Programming - Sequential File Read

SELECT with FETCH - COBOL Program

Processing Multiple Rows

Multi-row SELECT TABLE

RESULT TABLE

Retrieve One row at a time

DECLARE CURSOR

EXEC SQL DECLARE CUREMP01 CURSOR FOR SELECT EMPNO, SEX, AGE, NAME FROM EMPLOYEE WHERE JOIN_DATE > : WS-JOIN-DATE END-EXEC.

OPEN

EXEC SQL OPEN CURSOR-NAME END EXEC.

101 02/01/1996 102 11/01/1995 103 15/03/1995 104 15/02/1996

EXEC SQL OPEN CUR END-EXEC

105 15/02/1996 101 02/01/1996 102 05/04/1996

FETCH
Syntax : EXEC SQL FETCH CURSOR-NAME INTO :host-variable1, :host-variable2, .. END-EXEC.

Example : EXEC SQL FETCH CUREMP01 INTO :WS-EMPNO, :WS-SEX, :WS-AGE END-EXEC.

End - of - Data Processing

RESULT TABLE
105 101 102 15/02/1996 02/01/1996 04/05/1996

WS-EMPNO WS-JOIN-DATE

SQLCODE

FETCH

105 101 102 102

02/05/1996 01/02/1996 04/05/1996 04/05/1996

0 0 0 100

Handling End-of-data
MAIN-PARA.

MOVE 01-01-1996 TO WS-JOIN-DATE.

PERFORM OPEN-PARA

PERFORM FETCH-PARA

PERFORM UNTIL SQLCODE = 100

PERFORM PROCESS-ROW PERFORM END-PERFORM

FETCH-PARA

....

CLOSE

Syntax : EXEC SQL

CLOSE cursor-name
END-EXEC. Example : EXEC SQL

CLOSE CUREMP01
END-EXEC.

EMBEDDED SQL PROGRAM -CURSOR

EMBEDDED SQL PROGRAM -CURSOR

EMBEDDED SQL PROGRAM -CURSOR

EMBEDDED SQL PROGRAM -CURSOR

PRE-COMPILE

BIND

OUTPUT

Output

Output

OUTPUT

FOR UPDATE OF
EXEC SQL

DECLARE CUREMP01 CURSOR FOR

SELECT EMPNO, AGE FROM EMPLOYEE FOR UPDATE OF SALARY END-EXEC.

UPDATE WHERE CURRENT OF

PERFORM FETCH-PARA PERFORM UNTIL SQLCODE =100 .. EXEC SQL UPDATE EMPLOYEE SET SALARY = :WS-SALARY

WHERE CURRENT OF CUREMPO1

END-EXEC PERFORM FETCH-PARA END-PERFORM

DELETE WHERE CURRENT OF

PERFORM OPEN-PARA
PERFORM FETCH-PARA. PERFORM UNTIL SQLCODE =100 EXEC SQL DELETE FROM EMPLOYEE WHERE CURRENT OF CUREMP01 END-EXEC PERFORM FETCH-PARA

END-PERFORM.

DECLARE WITH HOLD

EXEC SQL DECLARE cursor-name CURSOR WITH HOLD FOR SELECT column-name-list FROM table-name WHERE search-condition FOR UPDATE OF column-name END-EXEC.

Dynamic SQL
Statement: Acquired during execution,and function can vary and can be done to different tables and columns

Bind: One single statement

At statement execution Access strategy not saved..

Static SQL vs. Dynamic SQL

Statement: Static : Object and Action known

Dynamic : Object or Action not known Bind: Static Dynamic : is bound once : is bound every time.

Coding Dynamic SQL

Translate the input data into a SQL statement Prepare the SQL statement to execute Execute the SQL statement Process the information returned Handle SQL return codes

EXECUTE IMMEDIATE

Prepare and Execute the SQL statement present in the COBOL host variable WS-STMT MOVE DELETE FROM EMPLOYEE TO WS-STMT-TEXT MOVE 50 EXEC SQL EXECUTE IMMEDIATE :WS-STMT END-EXEC. TO WS-STMT-LEN

Parameters Markers

Dynamic SQL statements cannot use Host Variables

Parameter Markers are used to substitute a parameter

DELETE FROM EMPLOYEE WHERE EMPNO = ?

Prepare and Execute

PREPARE once EXEC SQL PREPARE S1 FROM :WS-STMT EXEC-EXEC. EXECUTE many times EXEC SQL

PREPARE S1 FROM :WS-EMPNO

EXEC-EXEC.

DB2 LOCKS

Overview

DB2 LOCKS
DB2 LOCKS * * *
Are managed by IMS Resource Lock Manager, IRLM Are used to control Concurrency and guarantee the integrity of each users data Prevent access to uncommitted data

CONCURRENCY CONTROL

Concurrency is a situation wherein DB2 allows more than one application program to access same data at the same time.
Concurrency needs to be controlled due to many undesirable effects such as i LOST UPDATES (or Double Updating),

ii
iii

ACCESS TO UNCOMMITTED DATA

UNREPEATABLE READS

DB2 makes up use of locks to control Concurrency

Concurrency Control Problem

Lost Updates

Without Concurrency control two application programs X and Y

Both read the same Row from the table Both supply new values for one of its columns based on what they read.

PROBLEM:

If X updates the row with its value and then Y updates the same row, Xs updates are lost

Access To Uncommitted Data

Without concurrency control

Application program X might update a value in the table

Application program Y might read that value before it was committed
PROBLEM

If Xs value is not later committed but backed out, then Ys calculations are based on uncommitted and presumably invalid data

Concurrency control Problem

Unrepeatable Reads
Without concurrency control

An application program X reads a row form the table and goes on to process other SQL statements. X again reads the row it read the first time and must find the same values in it.

PROBLEM

Without Concurrency control, application program Y could have changed the same row between two reads of program X.

FACTORS CONTROLLING LOCKING

Overview

FACTORS CONTROLLING LOCKING

The Object Of A Lock

Is the resource being locked by DB2

Can be databases, table spaces, tables, tablespace page, index, indexspace, page/subpage and plans

The Size Of A Lock

Determines how large a portion of the resource is being locked The LOCKSIZE parameter of CREATE TABLESPACE statement defines the locking protocol DB2 being locking levels are:

Tablespace locking

Page locking

The Duration Of A Lock

Determines the length of time the lock is held Varies according to when the lock is acquired and when it is released

FACTORS CONTROLLING LOCKING

Duration Of Tablespace Locks

The duration of any tablespace lock held by an application program can be controlled by means of ACQUIRE and RELEASE parameters of BIND ACQUIRE specifies when resources should be allocated, I.e., when the locks should be acquired
RELEASE specifies when resources should be deallocated, I.e., when the locks should e freed or released. Both the above parameters have two sub-parameters:

For ACQUIRE
USE ALLOCATE

For RELEASE
COMMIT DEALLOCATE

Duration Of Page Locks

ISOLATION parameters controls the PAGE level locks

Bind Isolation Parameters

Influence the initial lock chosen by DB2 especially its duration Ensure data consistency, an application program can be bound with any of the two values Page lock is held only while the cursor is positioned on at page. When the cursor moves to another page, the lock is leased. When a page is locked concurrent application programs cannot update or delete a row of locked page If an application program updates or deletes data, the lock is held until the data is committed Applies only to data is read. All changed data remains locked until COMMIT or ROLLBACK

Cursor Stability

Duration Of Page Locks

Repeatable Read (RR)

Page lock is held until the next commit point.

With this option if the application program returns to the same page and reads the same row again, the values would remain the same It means that pages containing all the rows retrieved as well as the pages containing the current row are locked. Concurrent application programs cannot insert, update,or delete a row in the locked page until the unit of recovery completes.

Modes of Locks for Table or Tablespace

S(SHARE)

Lock owner and concurrent application programs may only.

Lock owner can read for updating it; TABLE SHARING Concurrent application programs may only read data. Lock owner can only read or change data in the table or tablespace; No other concurrent application programs can is allowed access in the Tablespace; Lock owner can only read data in the table or tablespace; Concurrent application programs can only read data in the locked page but can read and update data of other pages. the lock owner can read and update in the table ir tablespace

U(UPDATE)

X(EXCLUSIVE)
TABLE EXCLUSIVE IS(INDENT SHARE)

IX(INDENT EXCLUSIVE)

PAGE EXCLUSIVE : Concurrent application programs can only read and update data of other pages.