SUM with OUTFIL in One Step

Processing Order

When both SUM and OUTFIL are present, DFSORT runs: (1) Sort (and optional INREC), (2) SUM (collapse each group to one record), (3) OUTFIL (format and write to FNAMES=). So OUTFIL never sees the original detail records; it sees the summary records (one per group). SORTOUT also receives the summary records unless you use OUTFIL FNAMES=SORTOUT to replace what goes to SORTOUT with the OUTFIL output (e.g. summary records plus TRAILER1).

Typical Use: Summary Dataset and Report

A common pattern is: sort by key, SUM to get one record per group, then OUTFIL to write that summary to SORTOUT and to a report DD with a header and trailer:

1
2
3
4
5
6
7
  SORT FIELDS=(1,5,CH,A)
  SUM FIELDS=(20,6,PD)
  OUTFIL FNAMES=(SORTOUT,REPORT),
    HEADER1=(1:'Summary by Department',/,5:DATE=(MD4-)),
    OUTREC=(1,80),
    TRAILER1=(1:'Groups: ',COUNT=(M11,LENGTH=6),
      '  Grand total: ',TOTAL=(20,6,PD,LENGTH=12))

SORTOUT and REPORT both get the summary records (one per department) plus the header and trailer. COUNT= in TRAILER1 is the number of summary records (groups). TOTAL= sums the 6-byte PD field at 20 across those records—which is the same as the grand total of the original amount field because each summary record holds the group total.

Two Outputs: Summary Only vs Report

You can split outputs: one DD gets only the summary records (no header/trailer), another gets the report (summary records + HEADER1 + TRAILER1). Use two OUTFIL statements with different FNAMES= and different options:

1
2
3
4
5
  OUTFIL FNAMES=SUMMARY,OUTREC=(1,80)
  OUTFIL FNAMES=REPORT,
    OUTREC=(1,80),
    HEADER1=(1:'Report',/),
    TRAILER1=(1:'Total groups: ',COUNT=(M11,LENGTH=6))

SUMMARY gets 80-byte summary records only. REPORT gets the same records plus a one-line header and a trailer with the count of groups.

What TRAILER1 Count and Total Mean

After SUM, the record stream has one record per group. So COUNT= in TRAILER1 is the number of groups (summary records), not the original input record count. TOTAL= sums the specified field across the summary records. If that field in the summary record is the group total (from SUM), then the TRAILER1 total is the grand total across all groups. So for “grand total of amount,” TOTAL= is correct; for “total number of original records,” you would need a different approach (e.g. a separate step without SUM that only counts).

Explain It Like I'm Five

First we sort the cards and then we squish each pile (group) into one card that has the total for that pile. That’s SUM. Then we take those squished cards and we can put them in one box (SORTOUT) and also write a little title on top and a line at the bottom that says “how many piles” and “big total.” That’s OUTFIL. So one step: sorted, squished, and then either just the squished cards or the squished cards with a header and footer.

Exercises

1. Write control statements that sort by 1,5 CH A, sum 20,6 PD, and write summary records to SORTOUT and to REPORT with TRAILER1 showing count and total of the summed field.
2. Why does COUNT= in TRAILER1 give the number of groups (not original records) when SUM is used?
3. If you need both a file of summary records and a file of all detail records with a grand total trailer, why might you need two steps?