Example 6 - Paired and unpaired F1/F2 records (FILL method)

//JKE6  EXEC  PGM=SORT
//SYSOUT    DD  SYSOUT=*
//SORTJNF1 DD DSN=FIRST.FILE,DISP=SHR
//SORTJNF2 DD DSN=SECOND.FILE,DISP=SHR
//F1ONLY DD SYSOUT=*
//F2ONLY DD SYSOUT=*
//BOTH DD SYSOUT=*
//SYSIN    DD    *
* Control statements for JOINKEYS application
  JOINKEYS FILE=F1,FIELDS=(1,10,A),SORTED,NOSEQCK
  JOINKEYS FILE=F2,FIELDS=(7,10,A),SORTED,NOSEQCK
  JOIN UNPAIRED,F1,F2
  REFORMAT FIELDS=(F1:1,14,F2:1,20),FILL=C'$'
* Control statements for main task (joined records)
  OPTION COPY
  OUTFIL FNAMES=F1ONLY,INCLUDE=(15,1,CH,EQ,C'$'),
    BUILD=(1,14)
  OUTFIL FNAMES=F2ONLY,INCLUDE=(1,1,CH,EQ,C'$'),
    BUILD=(15,20)
  OUTFIL FNAMES=BOTH,INCLUDE=(15,1,CH,NE,C'$',AND,1,1,CH,NE,C'$'),
    BUILD=(1,14,/,15,20)
/*

This example illustrates an alternate way to create three output files; with paired F1/F2 records, unpaired F1 records and unpaired F2 records. It produces the same output as Example 5. Whereas Example 5 uses an indicator in one position to determine where the key was found, Example 6 uses a fill character in different positions to determine where the key was found. Another drawback of the Example 6 method is that you must use a FILL character that does not appear in either input record. The Explanation for Example 6 is the same as for Example 5 up to the REFORMAT statement and then it differs as follows:

The REFORMAT statement defines the fields to be extracted for the joined records in the order in which they are to appear. FIELDS=(F1:1,14,F2:1,20) tells DFSORT to create the joined records as follows:

Joined Record Positions   Extracted from
-----------------------   -----------------
1-14                      F1 positions 1-14
15-34                     F2 positions 1-20

FILL=C'$' tells DFSORT to use $ as the fill character for the F1 field in an unpaired F2 record and for the F2 field in an unpaired F1 record. We use the FILL character to identify the unpaired records from each file; when used for this purpose, the default of FILL=X'40' is usually not a good choice. You must select a FILL character that will not appear in either input file.

The OPTION COPY statement tells DFSORT to copy the joined records. The OUTFIL statements use the presence or absence of the $ fill character in certain positions to determine where to find the F1 or F2 fields in the joined records and where to write the fields (F1ONLY, F2ONLY or BOTH).

Conceptually, JOINKEYS application processing proceeds as follows:

Subtask1 copies the SORTJNF1 (F1) records as directed by the JOINKEYS statement. As a result, it copies the unchanged SORTJNF1 records to the main task.
Subtask2 copies the SORTJNF2 (F2) records as directed by the JOINKEYS statement. As a result, it copies the unchanged SORTJNF2 records to the main task.
The main task joins the records passed from subtask1 and subtask2 as directed by the specified JOINKEYS, JOIN and REFORMAT statements, resulting in the following joined records (paired and unpaired):
```
Carrie    F101No    Carrie    F201
David     F102$$$$$$$$$$$$$$$$$$$$
Frank     F103$$$$$$$$$$$$$$$$$$$$
Holly     F104Yes   Holly     F202
$$$$$$$$$$$$$$Yes   Karen     F203
$$$$$$$$$$$$$$No    Sri Hari  F204
Vicky     F105Yes   Vicky     F205
```
For F1 records without a match in F2 (for example, the F102 record), the F2 field is filled with the FILL character. For F2 records without a match in F1 (for example, the F203 record), the F1 field is filled with the FILL character. For F1 records with a match in F2 (for example, the F101 and F201 records), no FILL characters are used.
The first OUTFIL statement finds records with the FILL character in position 15. These are the F1 records without a match in F2. The F1 field is in positions 1-14 of the joined record, so those positions are written to the F1ONLY file. Thus, F1ONLY contains these records:
```
David     F102
Frank     F103
```
The second OUTFIL statement finds records with the FILL character in position 1. These are the F2 records without a match in F1. The F2 field is in positions 15-34 of the joined record, so those positions are written to the F2ONLY file. Thus, F2ONLY contains these records:
```
Yes   Karen     F203  
No    Sri Hari  F204
```
The third OUTFIL statement finds records without the FILL character in position 1 or position 15. These are the F1 and F2 records with a match. The F1 field is in positions 1-14 of the joined record and the F2 field is in positions 15-34 of the joined record, so each joined record is split into those two records and written to the BOTH file. The shorter F1 record is padded with blanks on the right to the length of the F2 record. Thus, BOTH contains these records:
```
Carrie    F101
No    Carrie    F201
Holly     F104
Yes   Holly     F202
Vicky     F105
Yes   Vicky     F205
```

Note: If you only want one record per key in BOTH, you can have the BUILD for FNAMES=BOTH specify the positions for just that record. For example, BUILD=(1,14) for the F1 records or BUILD=(15,20) for the F2 records.