Reminiscence...

A refreshing session on IMS database concepts

Anusha Battu

IMS – INFORMATION MANAGEMENT SYSTEM

Developed in 1968 by IBM

Data Language-I (DL/I) was developed as a interface between application code and data in IMS database

It supports Hierarchical DB structure

What is a hierarchical database…???

There is always a basic or "root" segment. A segment may relate downwards to one or more subordinate

segments; these subordinates are referred to as "child" segments. A subordinate (child) segment may relate upward to only one superior

or "parent" segment. Follows INVERTED TREE STRUCTURE

Structure of IMS Database

B3

A2

D1B2

A1

C1

ROOT SEGMENT (PARENT)

TWIN SEGMENTS DEPENDENT SEGMENTS

-- Level 1

-- Level 2B1

SEGMENT OCCURRNCES

IMS terminologyEach box in the database hierarchy chart represents a SEGMENT

SEGMENT -smallest block of information that DL/I can access in the database.

Segment name -max of 8 chars

Field name -max of 8 chars Each segment contains one or more fields of information. Two types of fields

Key FieldSearch FieldKey field is used to sequence the database and is unique. It cannot be changed where as search field value can be change according to our prog requirement

FIELD - Field is the smallest unit of info in the database.

Segments

PARENT SEGMENT - segment that has one or more segments dependent on it.

These dependent segments must be on the level directly below the parent.

CHILD SEGMENT - segment that is directly dependent on some other segment.

A segment without parent is called as ROOT SEGMENT.

DEPENDENT SEGMENTS - all the segments under a particular segment of a database.

SEGMENT OCCURRENCE - particular instance of a segment type

TWIN - occurrences (2 or more) of a segment type under a parent

SIBLINGS - occurrences of different segment types under a parent

RECORD - one occurrence of a root segment along with all its dependent segments

Segment Format

Prefix Data Segment

Code

1 byte

Delete byte

1 byte

Counters and

Pointers4bytes perelement

Size field

2 bytes

Seq. (key)field

Data

Segment Code To identify each segment type stored in a database (1 to 255).

Delete byte To track the status of a deleted segment

Counters andPointers

Pointers consist of one or more addresses of segments pointed to by this segment. Counter information is used when logical relationships are defined.

Size field This field states the size of the segment, including the size field (2 bytes).

Seq. (key) The sequence field is often referred to as the key field.

Data To contain the actual data being stored in the database.

IMS DATABASE LIMITATIONS

15 levels 255 segments 1000 fields No limitation on number of segment occurrences.

Physical & Logical DatabaseA

B C

D E

A

C

E

PHYSICAL DATABASE Represents how the data is actually recorded or stored on the storage medium (usually disk). Database records, composed of the root segment and all of its dependent segments are stored in hierarchical sequence:Top to bottomLeft to rightPhysical Database is defined in the Data Base Description (DBD).

LOGICAL DATABASE Represents the sequence of segments from one or more physical databases required by an application program. These segments are combined to create a new hierarchical structure.The logical data structure is defined in the Program Specification Block (PSB).

DBD

The DBA initiates a process called DBDGEN (database description generator) to describe the physical structure of the database and to create a DBD (database description).

Defines the layout and characteristics of the database Name of the DBD Type and access method for the DBD (HIDAM, HDAM, etc.) DD Name for the database Device Type Block Size Name, parent, length of each segment type in the database. Name, length, starting position, and type of data for each sequence and

search field in each segment

Example for DBDGEN Control statements:

PRINT NOGENDBD NAME=VENDOR,ACCESS=HDAMDATASET DD1=VEND,DEVICE=3380SEGM NAME=VENSEG,PARENT=0,BYTES=10FIELD

NAME=(VENCODE,SEQ,U),BYTES=10,START=1,TYPE=CSEGM NAME=ITEMSEG,PARENT=VENSEG,BYTES=5FIELD

NAME=(ITEMCODE,SEQ,U),BYTES=5,START=1,TYPE=CSEGM NAME=LOCNSEG, PARENT=ITEMSEG, BYTES=9FIELD NAME=(LOCNCODE,SEQ),BYTES=3,START=1,TYPE=CFIELD NAME=ORDDATE,BYTES=6,START=4,TYPE=CDBDGENFINISHEND

Macros/Control statements

DBD: The DBD statement...• Names the database• Identifies the access method

DATASET:The DATASET statement specifies.• The DD name used in the JCL• Disk DEVICE type (e.g. 3380)• Block size (e.g. 4096)• Free space % (FRSPC)

Macros/Control statements

SEGM:There is one SEGM statement for each segment in the database. SEGM ... • names the segment (max 8 characters) • identifies the PARENT segment• establishes the length of the segmentNOTE: PARENT=0 signifies the root segment.

FIELD:FIELD statements are used to define KEY fields and SEARCH fields... • SEQ defines EMPID to be a KEY field• U defines EMPID to be a UNIQUE key• BYTES defines the field size• START defines starting position in the segment for this field• TYPE defines field type

Typ DescriptionCobol

Picture

CCharacter(Default) X

P Packed Decimal COMP-3

Z Zoned Decimal S9

H Half word Binary 9(4) COMP

F Full word Binary 9(8) COMP

PSB• The PSB is made up of one or more PCB's (Program Communication Block), and this

defines the logical structure a program will use.

• The PSBGEN process produces a load module stored in the PSBLIB library. These load modules are called PSB's.

• The PSB is a control block that defines the logical data structure a program will use.

• While many application programs can share the same PSB (but usually do not), programs can only use one PSB in a given application.

• The Program Communication Block defines which of the segments in the database the program is "sensitive" to (can access).

• The PCB also defines how the application program is allowed to process the segments.

• The Program Specifications Block (PSB) supplies IMS with a description of what database information (which database segments) may be accessed by the application program. Each PSB consists of one or more Program Communication Blocks (PCB).

Example:

PCB TYPE=DB,DBDNAME=EMPDBD,KEYLEN=16 SENSEG NAME=EMPDATA,PARENT=0,PROCOPT=K SENSEG NAME=WORKDATA,PARENT=EMPDATA,PROCOPT=G PSBGEN LANG=COBOL,PSBNAME=EMPPSB1 END

TYPE:Data type: CharacterLength: 4Valid values: DB, TP and GSAM.

DBDNAME: Name of the Database the program needs access to.Data type: CharacterLength: 8

KEYLEN: The length in bytes of the longest concatenated key for the segments requested in the PSB. Segment key length may be found in the DBD.

Data type: SMALLINT

SENSEG The name of the segments in the database your program needs access to.

NAME: Name of the segment sensible to the programData type: CharacterLength: 8

PARENT: Name of the Parent SegmentData type: CharacterLength: 8

PROCOPT: Defines how the segment may be accessed (G, I, R, D, A, K, L etc.)Data type: CharacterLength: 4

Valid PROCOPTs include:

G GET (READ)

I INSERT

R REPLACE

K Access to only the KEY of the segment

O ONLY: used with G to indicate that Get Hold calls are not allowed

P Position function (Path calls)

D DELETE

A All options (G, I, R, D)

L LOAD (initial load of data)

LANG: Programming language in which the application program is coded.Data type: CharacterLength: 5

PSBNAME: It gives the PSB a name (usually the same name as the application program name that will use the PSB).

IMS-COBOL application program:

Working Storage Section • Function codes • Segment search argument (SSA)• Input/output area

Linkage Section • program communication block (PCB mask)

Procedure Division

• Entry and exit statements • DL/I call statement; status code checks

Function Codes: A Function code is a 4-byte word used to tell DL/I what kind of call to make

CODE MEANING DESCRIPTION

GU Get-Unique To obtain a unique occurrence of a segment within a database.

GN Get-Next To obtain the next occurrence of a segment within a database.

GNP Get-Next-Within-Parent To obtain the next occurrence of a segment under a parent; retrieves

child segments under a parent sequentially.

GHU Get-Hold-Unique The GET-HOLD function codes identical

to the GET codes except that they must

be used for retrieval if you intend to ALTER the data being retrieved.

GHN Get-Hold-Next

GHNP Get-Hold-Next-Within-Parent

REPL REPLACE To replace an occurrence of a segment

DLET DELETE To delete an occurrence of a segment

ISRT INSERT To insert an occurrence of a segment into database

SSA:

It is an optional DL/I call parameter that qualifies a call by specifying a particular segment type or occurrence

In SEGMENT SEARCH ARGUMENTS (SSAs) you can provide:

• The name of the segment type you want. • A description of a specific segment. • One or more "command codes" to qualify your DL/I call even more.

I/O: Input/ Output area

• It is a standard record description in the WORKING STORAGE SECTION which hold database segments for manipulation.

• Must be equal to or greater than the length of the longest segment to be used By the program

PCB MASK: • It is used to transfer information about each call made to DL/I back to the program. • You must code one Database PCB MASK for each database your program will access. • The 01 level name should be unique for each database.

Example:WORKING STORAGE SECTIONLINKAGE SECTION 01 DB-PCB-1. 03 DBD-NAME PIC X (8). 03 SEG-LEVEL PIC X (2). 03 STATUS-CODE PIC X (2). 03 PROC-OPTIONS PIC X (4). 03 IMS-ADDRESS PIC X (4). 03 SEGMENT-NAME PIC X (8). 03 KEY-LENGTH PIC S9 (5) COMP. 03 NUM-SENS-SEGS PIC S9 (5) COMP. 03 KEY-FEEDBACKS.

05 SEG-KEYS PIC X (#)

The most important field in PCB mask is STATUS-CODE. This field informs you if the DL/I call is successful or not.

You must specify the longest possible concatenated key.

DPRD#G.CHGMAN.CPY.BASE0(DPCMASK)

ENTRY and EXIT statements:

It must be the first executable statement in the program. It must follow immediately after the title PROCEDURE DIVISION.

NOTE: List the names of the DB-PCB mask in the same order as they are defined in the PSB for your program.

For exit, you can use GOBACK (not STOP RUN).

DLI CALL

CALL 'CBLTDLI' USING FUNCTION-CODE, PCB-MASK, IO-AREA, SSA-1, SSA-2...

SSA

SSA When specified in a DL/I call is an optional DL/I CALL parameter that qualifies a call by specifying a particular segment type OR segment occurrence.

The SSA always follows the I/O area parameter

There may be 1 to 15 SSA’s specified in a call stmt.

The SSA’s must appear in hierarchical order by segment type.

The SSA contains the information necessary to obtain the segment you require.

This information could be –

■ The segment name. ■ The segment name and segment key. ■ The segment name and search field.

Types of SSA

1. Unqualified SSA: Searches for a specific segment type

An unqualified SSA is a segment search argument which specifies a segment name ONLY. The 9th position in the literal must be blank. This indicates an unqualified SSA.

01 UNQUAL-SSA. 02 SEG-NAME PICTURE X(08) VALUE '........'.02 FILLER PICTURE X VALUE ' '.

2. Qualified: Searches for a specific segment occurrence of a specific segment type.

Supplies the DL/I call with either a KEY or SEARCH FIELD value in addition to the SEGMENT NAME.

A left parenthesis '(' in position 9 or after a variable number of command codes signifies the beginning of a qualification statement.

The right ')' closing parenthesis signifies the end of the qualified SSA.

01 SSA-EMPLOYEE. 03 SEGMENT-NAME PIC X(8) VALUE 'EMPLOYEE'. 03 COMMAND-CODE PIC X(2) VALUE '*-'. 03 BEGIN-QUALIFY PIC X(1) VALUE '('. 03 KEYNAME PIC X(8) VALUE 'EMPKEY '. 03 OPERATOR PIC X(2) VALUE '='. 03 KEY-VALUE PIC X(11). 03 END-QUALIFY PIC X(1) VALUE ')'.

OPERATOR: Relational Operator

>(GT), < (LT), >=(GE), <=(LE), =EQ, ¬=(NE)

COMMAND CODE:

Command codes are designated by a ‘*’ immediately fallowed by one or more command codes.

The initial value of COMMAND CODE is '*-'. The '-' is a NULL COMMAND

Code FunctionC Use a concatenated key

D Retrieve a path

N Do NOT replace this segment

P Establish parentage at this level

F Access first segment occurrence

L Access last segment occurrence

U Maintain current position at this level

V Maintain current position at this and higher levels

- Null/ignore

FUNCTION CODES:GU: To obtain a unique occurrence of a segment within a database. Qualified SSA is used No SSA Root Segment

Blank SuccessfulGE Not FoundGB End of the DB

Example: The SSA is used to build the GU call. 01 SSA-1. 03 SEGMENT-NAME PIC X(8) VALUE 'SEGMENTA'. 03 COMMAND-CODE PIC X(2) VALUE '*-'. 03 BEGIN-QUALIFY PIC X(1) VALUE '('. 03 KEYNAME PIC X(8) VALUE 'AKEY '. 03 OPERATOR PIC X(2) VALUE ' ='. 03 KEY-VALUE PIC X(11) VALUE '12'. 03 END-QUALIFY PIC X(1) VALUE ')'.

GU SEGMENTA *-(AKEY = 12)

GN: To retrieve segments sequentially

Used with 1.No SSA 2.Unqualified SSA 3.Qualified SSA

Blank SuccessfulGE Not Found GB End of DB

Example:

GNP: Get Next within Parent is used to sequentially access dependent

segments of an established parent segment. The first thing you must do before using GNP is to establish

parentage. There are two ways:1. By issuing a successful GU or GN call.2. By using the P Command Code (Establish Parentage) with a GU,

GN, or GNP call.

Effected by GU, GN, ISRT or DLET

GP: Lowest Segment in the call is not under parent.

PATH CALL:

To retrieve a segment and all of its higher level segments without having to issue a call for each one.

For this you need to use the Command code ‘D’ which is called as Path Command Code used for Path Calls.

The Program Specification Block (PSB) for your program must specify an additional PROCOPT value (P) in order to make path calls.

GN SEGMENTA *D(AKEY =1) SEGMENTB *D(BKEY =2) SEGMENTD *-(DKEY =5)

REPL:

Should be preceded by GHU or GHN Update the I/O area No SSA N No Replace

AJ: If SSA AM: Replacing segments which doesn't have Replace sensitivityDA: Key is changedRX: Violation of Replace rules

Steps:1.GHU call2.Update I/O3.REPL call

INSERT: To 'load' records to an IMS database, you use the insert function code - ISRT - in a DL/I

call. To add new records or segments to an existing IMS database you use the same insert

function code - ISRT - in a DL/I call. So, the ISRT call is used in two modes,

LOADINSERT

LOAD Mode: In LOAD mode, only two processing options are allowed:L (LOAD) and LS (LOAD SEQUENTIAL).LS is usually preferred as it specifies that segments must be loaded in ascending sequence. L processing option means load the database, but not necessarily in sequential order.

INSERT Mode: In INSERT mode, processing options can be:I - insert new segments in sequential or random order. IS - add new segments only in

ascending sequence.

■ The segment to be inserted must first be built in the program IO-AREA.

■ Use SSAs to define the path to DL/I.

■ Using the 'D' COMMAND CODE to insert a path of segments.

■ Use an unqualified SSA at segment level(s) where the insertion will occur.

■ New segments added to an existing database are inserted based on the segment KEY field:

■ Segments with a KEY field can be inserted in KEY sequence.

■ Segments with no KEY or non-unique KEY will be inserted according to Insert Rules specified in the DBD.

STATUS CODES for ISRT:II: Tries to insert a segment that already existsGE: Could not find the path of the segments specified in our SSA’s.

DLET:

Just like REPL call, DLET call must be preceded by one of the three GET HOLD calls.

Deletes a segment along with all its dependent segments No SSA is used Only one situation to use SSA in DLET call

DA: Key is changedDJ: No GH callDX: Delete rule violation

Secondary Indexes What is a Secondary Index? Secondary Index is a alternate accessing path to access a segment

in IMS DB which in turn improves the response and faster access.

Why Secondary Indexes? For accessing a segment type based on the search field in a

dependent segment type of search field within the segment type so that we can avoid sequential scan of entire database

Allow you to process a segment type in a sequence other than the one defined by the segment's key.

Characteristics of Secondary Indexes A secondary index is in its own separate database so it can be processed as a separate

database and must use VSAM as its access method.

Are invisible to the application program. And program to use must code PROCSEQ= parameter in the PCB

If an application program needs to do processing using the regular processing sequence, PROCSEQ= is simply not coded.

If the application program needs to do processing using both the regular processing sequence and the secondary index, the application program's PSB must contain two PCBs, one with PROCSEQ= coded and one without.

There can be 32 secondary indexes for a segment type and a total of 1000 secondary indexes for a single database.

3 Types of Segment used for S. Ix.

Target Segment. It is the segment the application program needs to retrieve. Can be at any of 15 level.

Source Segment. contains the field that the pointer segment has as its key field. Data is copied from the source segment and put in the pointer segment's key field.

Pointer Segment. The pointer segment is contained in the secondary index database and is the only

type of segment in the secondary index database.

Notes:- Target and source can be same. Source is dependent and target DLI call with S.Ix Ptr Segment search Access target Segment

How a Secondary Index Is Stored Secondary Index are stored in a single VSAM KSDS if the key in the pointer segment is

unique.

If keys are not unique, an additional data set must be used (an ESDS) to store segments containing duplicate keys.

The pointer field at the beginning of the logical record exists only when the key in the data portion of the segment is not unique.

Pointer segments containing duplicate keys are stored in the ESDS in LIFO (last in, first out) sequence. The KSDS logical record is updated to point to the second duplicate.

the original pointer segment (the one in the KSDS) is retrieved last.

Format and Use of Fields in a Pointer Segment Like all segments, the pointer segment has a prefix and data portion. The prefix portion has a delete byte, and when direct rather than symbolic pointing is used, it

has the address of the target segment (4 bytes). All fields in the data portion of a pointer segment contain data taken from the source

segment (with the exception of user data).

Delete Byte: The delete byte is used by IMS™ to determine whether a segment has been deleted from the database.

Pointer Field: This field contains the RBA of the target segment.

The pointer field exists when direct pointing is specified for an index pointing

Constant Field: This field, when present, contains a 1-byte constant. The constant is used when more than one index is put in an index database.

Search Field: The data in the search field is the key of the pointer segment.

All data in the search field comes from data in the source segment.

As many as five fields from the source segment can be put in the search field. Need not be contiguous in the source segment but will be concatenated in the search filed. The data in the search field (the key) can be unique or non-unique.

Subsequence Field: The subsequence field, like the search field, contains from one to five fields of data from the source segment.

Subsequence fields are optional, and can be used if you have non-unique keys. When a subsequence field is used, the subsequence data is concatenated with the data in the

search field. These concatenated fields become the key of the pointer segment. The subsequence field can make non-unique keys unique. Max of 5 fields in the source segment can be used in this field. This field cannot be seen by application program.

Duplicate Data Field: The duplicate data field, like the search field, contains from one to five fields of data from the source segment.

Use duplicate data fields when you have applications that process the secondary index as a separate database.

This field cannot be seen by application program.

Concatenated Key Field: This field, when present, contains the concatenated key of the target segment.

field exists when symbolic pointing is used

Also, symbolic pointers generally require more space in the pointer segment.

User Data in Pointer Segments: You can include any user data in the data portion of a pointer segment by specifying a segment length long enough to hold it.

Like data in the subsequence and duplicate data fields, user data is never seen by an application program unless it is processing the secondary index as a separate database.

During reorganization of a database that uses secondary indexes, the secondary index database is rebuilt by IMS. During this process, all user data in the pointer segment is lost.

Making Keys Unique Using System Related Fields If you can’t make unique key using source fields

Using the /SX Operand:

FIELD NAME = /SXaaaaa In Source segment

You can code a FIELD statement with a NAME field that starts with /SX.

The /SX operand is specified in the SUBSEQ= operand in the XDFLD statement.

Creates RBA of the source segment as subsequent field ensuring that key is unique.

Using the /CK Operand:

FIELD NAME = /CKaaaaa In Source segment

You can code a FIELD statement with a NAME field that starts with /CK.

The /CK operand is specified in the SUBSEQ= operand in the XDFLD statement.

The /CK operand works like the /SX operand except that the concatenated key, rather than the RBA, of the source segment is used.

you can use a portion of the concatenated key of the source segment (if some portion will make the key unique) or all of the concatenated key using the parameters “BYTES=“ and “START=“.

Making key fields unique avoids the overhead of using an ESDS to hold duplicate keys.

How the Secondary Index is Maintained

IMS always begins index maintenance by building a pointer segment from information in the source segment that is being inserted, deleted, or replaced.

Inserting and Deleting a Source Segment: When a Source segment is inserted/deleted the corresponding pointer segment will be inserted/deleted by IMS.

Replacing a Source Segment: IMS determines what needs to be done by comparing the pointer segment it built (the new one) with the matching pointer segment in the secondary index (the old one).

If there is no change to the old pointer segment, IMS does not do anything. If the non-key data portion in the new pointer segment is different from the old one, the old

pointer segment is replaced. User data in the index pointer segment is preserved when the pointer segment is replaced.

If the key portion in the new pointer segment is different from the old one, the old pointer segment is deleted and the new pointer segment is inserted. User data is not preserved when the index pointer segment is deleted and a new one inserted.

Processing a Secondary Index as a Separate Database Because they are actual databases, secondary indexes can be processed independently.

2 Reasons to process Secondary as a separate database.

1. When you want small piece of data from database, you can put the data in the pointer segment itself. So that accessing the regular database can be avoided.

2. You can add to or change the user data portion of the pointer segment. The only way an application program can see user data or the contents of the duplicate data field is by processing the secondary index as a separate database.

The restrictions are as follows:

Segments cannot be inserted.

Segments can be deleted. Note, however, that deleted segments can make your secondary index invalid for use as an index.

The key field in the pointer segment (which consists of the search field, and if they exist, the constant and subsequence fields) cannot be replaced.

PROT, NOPROT: You can code the PROT/NOPROT operand in the DBD statement.

PROT: You can delete the pointer segment You can replace only the user data field in the pointer segment.

NOPROT: You can delete the pointer segment You can replace all the fields in the pointer segment except the key fields. PROT is the default

For using the secondary index as separate database Code a PCB for the application program. This PCB must reference the DBD for the secondary

index. The SSAs must use the complete key of the pointer segment as the qualifier. The PCB key feedback area in the application program will contain the entire key field.

How to Specify Use of Secondary Indexing in the DBD

Ex:

Where EDUC: Actual databaseSINDX: Secondary index database

DBD stmt for Actual database

Parameters in DBD stmt for Actual database:

An LCHILD and XDFLD statement are used to define the secondary index. These statements are coded after the SEGM statement for the target segment.

LCHILD: NAME= specifies the name of the secondary index SEGM statement and the name of the secondary

index database

PTR= parameter is always PTR=INDX when a secondary index is used.

XDFLD: Defines the content of the pointer segment and options used in the secondary index.

NAME= The name that can be used in the SSA to qualify a DL/I call on the secondary processing sequence.

SEGMENT= This parameter identifies the source segment this operand is omitted, the target segment is assumed to be the same segment as the source segment.

SRCH= Specifies the one to five fields from the source segment that are to be copied into the pointer segment's search field.

SUBSEQ= Specifies the one to five fields from the source segment that are to be copied into the pointer segment's subsequence field.

DDATA= Specifies the one to five fields from the source segment that are to be copied into the pointer segment's duplicate data field. This parameter is optional.

DBD stmt for S.Ix database

Parameters in DBD stmt for S.Ix database:

DBD statement:NAME= Specifies the name of the secondary index database. ACCESS= Parameter is always ACCESS=INDEX for the secondary index DBD.

SEGM statement:NAME= You choose name of the pointer segment. This is used in PCB when using the secondary index as a separate database.

FIELD statement:NAME= Parameter specifies the sequence field of the secondary index. the sequence field is composed of both the search and subsequence field data

LCHILD statement:NAME= Specifies the name of the target segment and target databaseINDEX= parameter has the name on the XDFLD statement in the target database. PTR= SNGL: Direct address pointer SYMB : Symbolic pointer

How SSAs with Secondary Indexes works ???

Use the name of an indexed field in your SSAs. IMS goes directly to the secondary index and finds the pointer segment with the value you

specify. Then IMS locates the segment that the index segment points to in the database and returns the

segment to your program.

Guidelines to use: Define the indexed field in the DBD for the primary database in the XDFLD statement during DBD

generation. Use the name that was given on the XDFLD statement as the field name in the qualification

statement. Specify the secondary index as the processing sequence during PSB generation. Do this by

specifying the name of the secondary index database on the PROCSEQ parameter on the PCB during PSB generation.

Can use two AND operators to connect the qualification statements: 1) ‘*’ or ‘&’, 2) #

Dependent AND:

Independent AND:

GU PATIENT(XILLNAME=TONSILITIS#XILLNAME=STREPTHRT)

GU PATIENT (XBILLING>=00500*XBILLING<=01000)

What DL/I Returns with a Secondary Index ???

Instead of giving you the key for the segment you requested, IMS gives you the search portion of the key of the secondary index (the key for the segment in the INDEX database).

I/O area looks just as it would if you had not used secondary indexing.

Status Codes for Secondary Indexes:-

NI: Try to insert or replace a segment that contains a secondary index source field that is a duplicate of one that is already reflected in the secondary index,

NE: Finds no pointer segment for an existing source segment

GSAM Databases

GSAM enables IMS™ batch application programs and BMPs to access a sequential data set as a simple database.

A GSAM database is a z/OS data set record that is defined as a database record.

The record is handled as one unit

it contains no segments or fields and the structure is not hierarchic.

To process a GSAM database, an application program issues calls similar to the ones it issues to process a full-function database.

The program can read data sequentially from a GSAM database, and it can send output to a GSAM database.

An IMS program can retrieve records and add records to the end of the GSAM database, but the program cannot delete or replace records in the database.

You use separate calls to access GSAM databases.

GSAM DBD generation During DBD generation for a GSAM database, you specify:

One data set group The ddname of an input data set that is used when an application retrieves data from

the database The ddname of an output data set that is used when loading the database

You cannot specify:

SEGM and FIELD statements The use of logical or index relationships between segments

DBD NAME=DDBBT003,ACCESS=(GSAM,BSAM),DATASET DD1=DDBBT003, RECFM=FB DBDGEN FINISH END

GSAM PCB Statement

PCB TYPE=GSAM,DBDNAME=DDBBT003, PROCOPT=G

TYPE=GSAM Is a required keyword parameter for all GSAM database PCBs

DBDNAME= Is a required keyword parameter for the name that specifies the GSAM DBD to be used as the

primary source of data set description

PROCOPT= the processing options on the data set declared in this PCB that can be used in an associated

application program.

Valid PROCOPTs:

G, GS, L and LS

Retrieving and Inserting GSAM Records To retrieve GSAM records sequentially, use the GN call.

To process the whole database, issue the GN call until you get a GB status code in the GSAM PCB.

GSAM automatically closes the database when you reach the end of it.

To create a new data set or to add new records to the end of the database, use the ISRT call.

GSAM adds the records sequentially in the order in which you supply them.

You can retrieve records directly from a GSAM database, but you must supply the record's address.

To do this, use a record search argument (RSA). An RSA is similar to an SSA, but it contains the exact address of the record that you want to retrieve

Explicitly Opening and Closing a GSAM Database IMS opens the GSAM data set when the first call is made and closes the data set when

the application program terminates.

However, explicit OPEN and CLSE calls are useful in the following two situations:

If the application program loads a GSAM data set, and then in the same step reads the data set using GSAM (for example, to sort the data set). The application program should issue the GSAM CLSE call after the load is complete.

If the GSAM data set is an output data set, and it is possible that when the program executes it does not make GSAM ISRT calls. A data set is not created. Subsequent attempts to read the nonexistent data set (using GSAM or not) will likely result in an error. To avoid this situation, explicitly open the data set. DL/I closes the data set when the step terminates. Closing the data set prevents the possibility of attempting to read an empty data set.

The explicit OPEN or CLSE call need not include an I/O area parameter.

Depending on the processing option of the PCB, the data set is opened for input or output.

GSAM Record Formats GSAM records are non keyed.

For variable-length records you must include the record length as the first 2 bytes of the record.

If you use undefined-length records, record length is passed between your program and GSAM in the 4-byte field that follows the key feedback area of the GSAM DB PCB.

When you issue an ISRT call, supply the length.

When you issue a GN or GU call, GSAM places the length of the returned record in this field

The advantage of using undefined-length records is that you do not need to include the record length at the beginning of the record, and records do not need to be of fixed length.

The length of any record must be less than or equal to the block size (BLKSIZE) and greater than 11 bytes (an MVS convention).

GSAM I/O Areas The format of the I/O area depends on whether the record is fixed-length, undefined-

length (valid only for BSAM), or variable-length. For each kind of record, you have the option of using control characters.

The formats of an I/O area for fixed-length or undefined-length records are:

With no control characters, the I/O area contains only data. The data begins in byte 0. With control characters, the control characters are in byte 0 and the data begins in

byte 1.

The formats for variable-length records

Without control characters, bytes 0 and 1 contain the 2-byte length field, and the data begins in byte 2.

With control characters, bytes 0 and 1 still contain the length field, but byte 2 contains the control characters, and the data starts in byte 3.

PCB Masks for GSAM Databases GSAM describes the results of those calls in a GSAM DB PCB.

Some of the fields are different from DB PCB masks, and the GSAM DB PCB has one field that the other PCBs do not.

Because GSAM is not a hierarchic database, some fields in a PCB mask for other IMS databases do not have meanings in a GSAM PCB mask.

01 DB-PCB-1. 03 DBD-NAME PIC X (8). 03 SEG-LEVEL PIC X (2). 03 STATUS-CODE PIC X (2). 03 PROC-OPTIONS PIC X (4). 03 IMS-ADDRESS PIC X (4). 03 SEGMENT-NAME PIC X (8). 03 KEY-LENGTH PIC S9 (5) COMP. 03 NUM-SENS-SEGS PIC S9 (5) COMP. 03 KEY-FEEDBACKS PIC X (8). 03 UNDEFINED-LENGTH PIC X (4).

Database Name. The name of the GSAM DBD. This field is 8 bytes and contains character data.

Segment Level Number. Not used by GSAM, but you must code it. It is 2 bytes.

Status Code. IMS places a two-character status code in this field after each call to a GSAM database.

Processing Options. This is a 4-byte field containing a code that tells IMS the types of calls this program can issue.

Reserved for IMS. This 4-byte field is used by IMS for internal linkage.

Segment Name. This field is not used by GSAM, but it must be coded as part of the GSAM DB PCB mask. It is 8 bytes.

Length of Key Feedback Area and Undefined-Length Records Area. This is a 4-byte field that contains the decimal value of 12. This is the sum of the lengths of the key feedback and the undefined-length record areas described below.

Number of Sensitive Segments. This field is not used by GSAM, but it should be coded as part of the GSAM DB PCB mask. This field is 4 bytes.

Key Feedback Area. After a successful retrieval call, GSAM places the address of the record that is returned to your program in this field. This is called a record search argument (RSA). This field is 8 bytes.

Undefined-Length Records Area. If you use undefined-length records (RECFM=U), the length in binary of the record you are processing is passed between your program and GSAM in this field. This field is 4 bytes long.

When you issue a GU or GN call, GSAM places the binary length of the retrieved record in this field. When you issue an ISRT call, put the binary length of the record you are inserting in this field before issuing the

ISRT call.

GSAM Status Codes AF

GSAM detected a variable-length record with an invalid format. Terminate your program.

AH You have not supplied an RSA for a GU call.

AI There has been a data management OPEN error.

AJ One of the parameters on the RSA that you supplied is invalid.

AM You have issued an invalid request against a GSAM database.

AO An I/O error occurred when the data set was accessed or closed.

GB You reached the end of the database, and GSAM has closed the database. The next position is the

beginning of the database.

IX You issued an ISRT call after receiving an AI or AO status code. Terminate your program.

Origin of GSAM Data Set Characteristics For an input data set, the record format (RECFM), logical record length (LRECL), and block size

(BLKSIZE) are based on the input data set label.

If this information is not provided by a data set label, the DD statement or the DBD specifications are used. The DD statement has priority.

Record Format:

Using the RECFM= parameter of the DATASET statement of DBDGEN, or DCB=RECFM=xxx on the DD statement. If the record format is specified on both, the DD statement has priority. V, VB, F, FB, U

Logical Record Length:

using the RECORD= parameter of the DATASET statement of DBDGEN, or DCB=LRECL=xxx on the DD statement. If the logical record is specified on both, the DD statement has priority.

Block Size:

Using the BLOCK= or SIZE= parameter of the DATASET statement of DBDGEN, or use DCB=BLKSIZE=xxx on the DD statement. If block size is specified on both, the DD statement has priority.

DD Statement DISP Parameter for GSAM Data Sets The DD statement DISP parameter varies, depending on whether you are

creating input or output data sets and how you plan to use the data sets:

For input data sets, use DISP=OLD.

For output data sets, you have a number of options:

If you are creating an output data set allocated by the DD statement, use DISP=NEW.

To create an output data set that was previously cataloged, but is now empty, use DISP=MOD.

When restarting the step, use DISP=OLD.

Finally, to add new records to the end of an existing data set, use DISP=MOD.

Suggested Method for Specifying GSAM Data Set Attributes

When specifying GSAM data set attributes:

On the DBD, specify RECFM. (It is required.)

On the DATASET statement, specify the logical record length using RECORD=.

On the DD statement, do not specify LRECL, RECFM, or BLKSIZE. The system determines block size, with the exception of RECFM=U. The system determines logical record length from the DBD.

For the PSB, specify PROCOPT=LS for output and GS for input. If you include S, GSAM uses multiple buffers instead of a single buffer for improved performance.

Use of Concatenated Data Sets by GSAM

GSAM can use concatenated data sets, which may be on unlike device types, such as DASD and tape, or on different DASD devices.

Logical record lengths and block sizes can differ.

The maximum number of concatenated data sets for a single DD statement is 255.

The number of buffers determined for the first of the concatenated data sets is used for all succeeding data sets. Generation data groups can result in concatenated data sets.

Database Recovery

Making sure that problems with the system have as little impact on the integrity of the DB and the efficiency of the operations systems as possible.

Abnormal Termination Routine:

When an application program fails, DL/I intervenes to contain the damage that’s done to the DB. It does this by invoking Abnormal Termination Routine.

ATR makes sure that the DB datasets are properly closed and their catalog entries are properly updated too.

In addition to that, DL/I log the program was using is also properly closed.

Although DL/I terminates your abending program in an orderly way, its function is to prevent additional database damage, not to correct damage that’s already been done.

Just because the ATR is executed does not insure that the DB in use is accurate.

Its necessary to back out changes made by the abending program, correct the error, then rerun the program. To do this DL/I log is required.

Logging

DL/I records all the changes a program makes to its DB in a special file called a Log that can reside on tape or DASD.

When a program changes a segment, DL/I logs both a Before Image and After Image of it which are used to restore a DB to its proper condition in case of program or system failure.

DL/I takes care of logging. In fact, logging often extends beyond your program’s view of the DB.

Recovery:

There are 2 types of Recovery

1. Forward Recovery2. Backward Recovery

Forward Recovery

Change data for a period of time is accumulated, then applied to a copy of the database as is existed before the changes began.

Forward recovery is normally used only when a DB is physically damaged in some way.

For ex, if a device error makes the current version of the DB inaccessible, forward recovery is used to restore it.

There are 2 requirements for this

1. Old copy of the DB has to be available2. All the changes posted to the DB since the copy was made must be available.

Dataset Image Copy Utility or Database Image Copy Utility is used to meet the first requirement.

Database Change accumulation Utility: Periodically combine all logs in case they have to be used to recover the DB.

The O/P of this program is called an accumulated change log.

In situations like a head crash on the DSAD that contains the DB dataset, it has to be forward recovered.

Forward Recovery can be a time consuming and awkward process.

Forward Recovery

Backward Recovery

Also called as Back out.

It’s a simple and more appropriate for most situations.

It used the DB at the time of failure and reverses all the changes made to it since the program began or since the program issued its last checkpoint.

During Back out, log records for the program are read backwards and their effects are reversed in the DB.

When the back out is complete, the databases are in the same state they were in before the failure.

Back out is appropriate when an application ends in a controlled fashion.

Recovery proceeds back to either the start of the program or to the last check point.

Check pointing

Also called as Synchronization point, Sync point, Commit point and point of integrity.

A point in the execution of the program where the DB changes are considered complete and accurate.

Database changes made before the most recent check point are not reversed by backward recovery.

DB changes logged after the most recent check point are not applied to an IC of DB during Forward Recovery.

So, Whenever the BR or FR is used, the DB is restored to its condition at the most recent check point when the recovery program completes.

Default for the batch programs is that the CP is the beginning of the program.

Basic Check pointing

To use check pointing, you code your program so that it periodically issues checkpoint calls.

Most programs issue check points at intervals like every 100 or 1000 transactions.

And some applications issue check points based on elapsed time, perhaps every 10 or 15 minutes.

CALL CBLTDLI USING DLI-CHKP I-O-PCB-MASK CHECKPOINT-ID.

For CHKP call under IMS, you supply the name of a special PCB called the I/O PCB.

Its normally used for Data communication programs. Here we use only the Status-Code field of the I-O-PCB.

01 I-O-PCB-MASK. 05 FILELR PIC X(10). 05 I-O-PCB-STATUS-CODE PIC XX.

You must list the I-O PCB as the first PCB in the Procedure division.

In Checkpoint-ID field, your program places a value that identifies the Checkpoint record.

Then, During Recovery, the operations team can use the checkpoint-ID to restart the program.

Symbolic Checkpointing

It is similar to basic checkpointing in that you use a CHKP call to write checkpoint records to a DL/I log.

But the symbolic checkpoint along with extended restart, provide an advantage: they let you store program data along with the checkpoint records and retrieve that data when its necessary to restart the program after a failure.

When you use symbolic checkpointing, the check point call begins with the same 3 parameters as the basic checkpoint call.

Then, you can code up to 7 pairs of field names to specify the working storage areas you want to have saved along with the checkpoint record.

In each pair, the first item is the name of the full word binary field (PIC S9(5) COMP) that contains the length of the data area to be saved, the second is the name of the data area itself.

CALL CBLTDLI USING DLI-CHKP I-O-PCB-MASK CHECKPOINT-ID AREA1-LENGTH AREA1

A program that uses extended restart should always issue an XRST call before it issues any other DL/I calls.

On XRST call, you list the same working storage fields you list on the CHKP call.

When the program is restarted, DL/I retrieves the values stored in the checkpoint record and restores the specified fields.

CALL CBLTDLI USING DLI-XRST I-O-PCB-MASK RESTART-WORK-AREA AREA1-LENGTH AREA1

If the program fails, the problem that caused the failure is corrected and the affected data bases are restored using forward or backward recovery.

Then, the program is restarted.

The operator supplies the last checkpoint-id on the PARM for the EXEC that invokes the program.

Then, the XRST call, which is the first call in the program, know the program should restart rather than begin a normal execution.

Symbolic CHKP and XRST with GSAM To checkpoint GSAM databases, use symbolic CHKP and XRST calls.

By using GSAM to read or write the data set, symbolic CHKP and XRST calls can be used to reposition the data set at the time of restart, enabling you to make your program restartable.

When you use an XRST call, IMS repositions GSAM databases for processing.

When restarting GSAM databases:

You cannot use temporary data sets with a symbolic CHKP or XRST call.

A SYSOUT data set at restart time may give duplicate output data.

Do not use DISP=MOD to add records to an existing tape data set.

Do not use DISP=DELETE or DISP=UNCATLG.

Thank you