Using IBM Application Development Tools for z/OS and OS/390

38 Using IBM Application Development Tools for z/OS and OS/390

Figure 2-20 Computer generated record

The default values for the fields are mostly fixed, as listed in Table 2-2.

Table 2-2 Some default values used in the example

2.1.9 How to split a single file into constituent record types

There may be times when you need to take one or more of the record types in a multi-record

file and segregate the records for additional processing. The recommended way to do this is

by using a DFSORT procedure. We show batch job streams, using both DFSORT and REXX

procedures, that take a file and split it into three record types. All other record types are

ignored.

The required DFSORT procedure would be similar to that shown in Example 2-8.

Example 2-8 DFSORT procedure

//*

//FILEMGR EXEC PGM=FMNMAIN

//STEPLIB DD DSN=FMN.V4R1M0.SFMNMOD1,DISP=SHR

//* DD DSN=IGY.SIGYCOMP,DISP=SHR

//REC01 DD DSN=REDBK2.TEST.SEQFILEA,DISP=OLD

//REC02 DD DSN=REDBK2.TEST.SEQFILEB,DISP=OLD

//REC03 DD DSN=REDBK2.TEST.SEQFILEC,DISP=OLD

//SYSPRINT DD SYSOUT=*

Field Type Default value

CUST-NM AN FX 'TEST ADT' repeat

KEYREC-DOT AN FX ‘x’

COMP-NM AN RO 'aaabbbcccdddeeefff'

DEC-NO-SHARES PD Init=10, Incr=7

BUY-FROM AN FX ‘Unknown’

BUY-FROM-NO AN RA '0123456789'

Chapter 2. Introducing File Manager 39

//FMNTSPRT DD SYSOUT=*

//SYSTERM DD SYSOUT=*

//SYSIN DD *

$$FILEM DSC DSNIN=REDBK2.TEST.SEQFILE,

$$FILEM POSITION=0,

$$FILEM DISP=MOD,

$$FILEM DSNOUT=REDBK2.TEST.SEQFILEX,

$$FILEM PROC=*

*FASTPROC

OUTFIL FNAMES=REC01,INCLUDE=(1,2,CH,EQ,C'01')

OUTFIL FNAMES=REC02,INCLUDE=(1,2,CH,EQ,C'02')

OUTFIL FNAMES=REC03,INCLUDE=(1,2,CH,EQ,C'03')

/+

The same requirement can be met by using an in-stream REXX routine to process the

records.The File Manager step of the batch job, shown in Example 2-9, illustrates this.

Example 2-9 File Manager step of batch job

//*

//FILEMGR EXEC PGM=FMNMAIN

//STEPLIB DD DSN=FMN.V4R1M0.SFMNMOD1,DISP=SHR

//* DD DSN=IGY.SIGYCOMP,DISP=SHR

//REC01 DD DSN=REDBK2.TEST.SEQFILEA,DISP=OLD

//REC02 DD DSN=REDBK2.TEST.SEQFILEB,DISP=OLD

//REC03 DD DSN=REDBK2.TEST.SEQFILEC,DISP=OLD

//SYSPRINT DD SYSOUT=*

//FMNTSPRT DD SYSOUT=*

//SYSTERM DD SYSOUT=*

//SYSIN DD *

$$FILEM DSC DSNIN=REDBK2.TEST.SEQFILE,

$$FILEM POSITION=0,

$$FILEM DISP=MOD,

$$FILEM DSNOUT=REDBK2.TEST.SEQFILEX,

$$FILEM PROC=*

DDNAME = 'REC' || FLD(1,2)

IF NCO(FLD(1,2),1,2,3) THEN DO

WRITE(DDNAME)

RETURN 'DROP'

END

/+

What is happening in these jobs

The file SEQFILE is the input file that contains multiple record types. The default output file

has the DD name SEQFILEX. Each of the record types we are interested in go into REC01,

REC02, or REC03.

The File Manager program keyword DSC is used to invoke the Data Set Copy function. The

input and output files are identified, and the keyword PROC=* is used to indicate that an

in-stream routine is being supplied.

The result is that all type 01 records end up in REC01, type 02 records go to REC02, type 03

records go to REC03, and all other record types go to the file EXTRA.

Let us review the report output

The key portion of the batch job’s output report is listed in Example 2-10.