For the TM/MFS Expert

DBUTIL - Interactive DB Calls

Chapter 8: DBD, PSB and MFS Statements

This chapter details the support IMS Option provides for the DBDGEN, PSBGEN and MFSGEN source statements.

See the chapters For the Database Administrator (DBA) and For the TM/MFS Expert in this guide for additional information.

8.1 Overview

The source statements are grouped in major categories of language control statements and definition statements. Language control statements, such as COPY, control the processing of the Gen utility. The definition statements define the DBD, PSB or MFS member. With each statement, the valid keywords or positional operands are listed with their possible values and a description of the support that IMS Option provides.

8.2 General Specifications

The format of the statements, positional operands and keywords are the same as defined by your IBM IMS/ESA manuals. This includes continuation rules, comments and labels except as noted here.

• Up to 39 continuation lines are supported. More than 40 lines for one statement generates a warning message and statement truncation.

• Literal values cannot span more than two lines

• All statements and keywords are assumed to be in upper case. Only literal values are processed correctly if lower case.

• In general, keywords or positional operands whose values must be numeric are edited to ensure the value is numeric. If not numeric, you see a warning message and the value is assumed to be zero. This can result in a subsequent warning or severe error message depending on the keyword or positional operand. All numeric values are limited to a maximum of eight digits.

• When the notes state the keyword and/or operand is edited, this indicates that at least some editing of the keyword, operand and/or value is performed. There is a complete list of the DBD, MFS and PSB warning and error messages in the section Error Messages in the online Reference help.

• All keyword operands which specify the string "*NAME" are eight character names which follow the convention for valid IBM mainframe member names. These names are always assumed to be valid.

• Except as noted, unrecognized keywords and/or values are ignored

• Except as noted, language and macro control statements are ignored. For example, PRINT, TITLE, EJECT, .AIF, .AGO.

• There is no edit recovery or MFS syntax error recovery. MFSGEN terminates when a severe error is encountered. MFSGEN does not try to recover from severe syntax errors.

8.3 DBDGEN

There are two areas, language control and statements.

8.3.1 DBDGEN Language Control - COPY

Any number of COPY statements can be defined. Nested COPY statements are not supported by DBDGEN. COPY statements are edited. ++INCLUDE and -INC statements are supported as if the ++INCLUDE or -INC statement was a COPY statement. The extensions provided by these two library management statements are not supported.

8.3.2 DBDGEN Statements

8.3.2.1 AREA

The AREA statement does not affect the processing of Fast Path DEDB databases and is not edited. All segments in a DEDB reside in the same area.

8.3.2.2 DATASET

The DATASET statement's effect on GSAM is summarized below. For more information about GSAM databases, see the sections GSAM Considerations in the chapter For the Database Administrator (DBA) and Unsupported DB Features in the chapter Product Summary.

8.3.2.3 DBD

Only one database can be defined in any one input data set.

8.3.2.4 DBDGEN

Supported and edited.

8.3.2.5 END

Supported and edited.

8.3.2.6 FIELD

A maximum of 1,000 FIELD statements can be listed in one DBD.

There is a restriction of 512 bytes on the length of the key that IMS Option uses internally to store each segment in a database. This length is calculated by summing the lengths of the longest sequence field at each level in the database and adding two bytes for each level in the database. For non-unique sequence fields (SEQ,M), add four bytes to its sequence field length for this calculation. For non-keyed segments, use four bytes as its length for this calculation. If this limitation is exceeded, you see DBDGEN error message number DBDUT0L11.

For example, a database with just three segments, a root and two children of the root. The root sequence field is 10 bytes long, the first child's sequence field is 15 bytes long and the second child's sequence field is 25 bytes long. The sum of the key lengths at each level is 10 plus 25 for a total of 35. Adding two bytes for each of the levels in the database gives an IMS Option key length of 39.

8.3.2.7 FINISH

Supported and edited.

8.3.2.8 LCHILD

The LCHILD statement is supported and edited for logical relationships and secondary indexes. The LCHILD statement for the primary index of a HIDAM database is not required and not edited.

A maximum of 32 secondary indexes can be defined in one DBD.

A maximum of 32 logical children can be defined in one DBD.

8.3.2.9 SEGM

A maximum of 255 SEGM statements can be defined in one DBD.

8.3.2.10 XDFLD

A maximum of 32 secondary indexes can be defined in one DBD.

8.4 PSBGEN

There are two areas, language control and statements.

8.4.1 PSBGEN Language Control - COPY

Any number of COPY statements can be defined. Nested COPY statements are not supported by PSBGEN. COPY statements are edited. ++INCLUDE and -INC statements are supported as if the ++INCLUDE or -INC statement was a COPY statement. The extensions provided by these two library management statements are not supported.

8.4.2 PSBGEN Statements

8.4.2.1 END

This statement is ignored.

8.4.2.2 PCB

A maximum of 500 PCBs can be defined in one PSB. There is no limit to the number of TP PCBs which can be defined, however, only the first 99 can be referenced by an application program.

8.4.2.2.1 PCB TYPE=TP

8.4.2.2.2 PCB TYPE=DB

8.4.2.2.3 PCB TYPE=GSAM

8.4.2.3 PSB

Only one PSB can be defined in a single input data set.

8.4.2.4 SENFLD

A maximum of 10,000 SENFLD statements can be defined for one PSB. A maximum of 255 SENFLD statements can be defined for one SENSEG.

8.4.2.5 SENSEG

A maximum of 3,000 SENSEGs can be defined in one PSB when using the PSBGEN3 format for PSBGEN. The maximum is 1,000 SENSEGs when using PSBGEN2 format.

A maximum of 255 SENSEGs can be defined for one PCB.

SENSEGs must be listed in the same physical sequence as they are defined in the DBD. See the chapter For the Database Administrator (DBA) for SENSEG and SEGM support when using logical and secondary data structures.

8.5 MFSGEN

8.5.1 MFSGEN Language Control

8.5.1.1 ALPHA

Partially supported and edited. You can find the complete description of ALPHA statement processing in the chapter For the TM/MFS Expert.

8.5.1.2 COPY

Any number of COPY statements can be defined. MFSGEN supports up to two levels of nesting for COPY statements. That is, a COPY member can contain a COPY statement whose member can also contain a COPY statement. COPY statements are edited. ++INCLUDE and -INC statements are supported as if the ++INCLUDE or -INC statement was a COPY statement. The extensions provided by these two library management statements are not supported.

8.5.1.3 DO and ENDDO

The description of DO and ENDDO expansion is described in the sections MFLD and DFLD later in this chapter.

8.5.1.4 EQU

The EQU statement is partially supported and edited. The complete description of EQU support is described in the chapter For the TM/MFS Expert.

8.5.1.6 STACK and UNSTACK

These statements are supported and edited. Multiple device types for the same MOD-name are processed differently than with IMS/ESA.

For reference, any stacks are written as data sets, not stored in memory and are named by using the stack ID, if present, with a .stk extension. The default when no stack ID is present is $mfsstak.stk. A maximum of 32 unique stack IDs are supported. These files are created in the Project's Listing folder.

These files are deleted according to the UNSTACK statement. If the last UNSTACK statement keeps the stack, these files remain on disk. The use of KEEP and DELETE is fully supported.

8.5.2 MID and MOD Statements

The following sections describe the MID and MOD statements for MFSGEN:

8.5.2.1 DO

8.5.2.2 ENDDO

Supported and edited.

8.5.2.3 LPAGE

There is a limit of 16,386 LPAGE statements which can be defined in a single message.

8.5.2.3.1 LPAGE for MSG TYPE=INPUT

8.5.2.3.2 LPAGE for MSG TYPE=OUTPUT

8.5.2.4 MFLD

There is a limit of 500 MFLDs which can be defined for a single input or output message. Only MFLDs which specify a DFLD-name are counted toward the limit of 500. That is, MFLD literals and MFLDs with only a LTH keyword are not counted. If the message contains LPAGE definitions, the limit is 500 MFLDs for each logical page. The limit of 500 is after any DO statement expansion.

8.5.2.4.1 MFLD for MSG TYPE=INPUT

8.5.2.4.2 MFLD for MSG TYPE=OUTPUT

8.5.2.5 MSG

There is a limit of 16,386 output messages which can be defined in a single input data set. The input message definition must be present in the same input data set as its corresponding output definition.

8.5.2.5.1 MSG TYPE=INPUT

8.5.2.5.2 MSG TYPE=OUTPUT

8.5.2.6 PASSWORD

The PASSWORD statement is not supported. However, any MFLDs defined after the PASSWORD and before the first SEG are logically removed from the input message. The mapping between the program and the input message is maintained.

8.5.2.7 SEG

There is a limit of 100 SEG statements which can be defined in any one input or output message.

8.5.2.7.1 For MSG TYPE=INPUT

8.5.2.7.2 For MSG TYPE=OUTPUT

8.5.2.8 MSGEND

Supported and edited.

8.5.3 Device Definition

The following sections describe the statements for controlling MFSGEN device definition:

8.5.3.1 DEV

Only 3270 display, 3270 printers and SCS1 printer devices are supported.

8.5.3.1.1 DEV for 3270 Display

8.5.3.1.2 DEV for 3270 Printer

8.5.3.2 DIV

8.5.3.3 DPAGE

There is a limit of 16,386 DPAGE statements which can be defined for any one FMT definition.

8.5.3.3.1 For 3270 Display

8.5.3.3.2 For 3270 Printer

8.5.3.4 DO

8.5.3.5 DFLD

There is a limit of 500 labelled DFLDs which can be defined for a single device page (DPAGE). If physical paging is used, the limit of 500 applies to each physical page. A labelled DFLD which is not referenced by an MFLD in the MID or MOD does not count toward this limit. The DFLD limit is after any DO statement iterations.

There is a cumulative limit of 19,996 bytes for a pool area to store MID literals, MID default literals, MOD literals, DFLD literals and PF key data. If this limit is exceeded, MFSGEN error message MFSUT0L17 is issued. The lengths are cumulative and are the lengths of the literals plus six bytes of overhead for each literal. Some of the 19,996 bytes is used for labelled DFLD data so the actual maximum size is somewhat less.

8.5.3.5.1 DFLD for 3270 Display

8.5.3.5.2 DFLD for 3270 Printer

8.5.3.6 ENDDO

Supported and edited.

8.5.3.7 FMT

There is no limit to the number of formats which can be defined in any one input data set. However, only those formats which are referenced by a MOD are processed.

8.5.3.8 FMTEND

Supported and edited.

8.5.3.9 TABLE and IF

Only one TABLE can be defined in a single input data set. Table processing is disabled if more than one TABLE is present. Only one IF statement is allowed within a TABLE. If more than one IF is present for a TABLE, the table is ignored. The only format allowed for the IF statement is:

IF DATA,^,' ',ENDMPPI

This statement generates an ENDMPPI command when the associated DFLD contains a non-blank value.

For the TM/MFS Expert

DBUTIL - Interactive DB Calls