Sort a dataset without having to produce another dataset

hallecodec · Posted: Fri Mar 02, 2007 2:44 pm

I just want to ask: is there a way to sort a dataset without having to produce another dataset? if there is, what jcl utility it would be?

ofer71 · Posted: Fri Mar 02, 2007 2:46 pm

For any of the common SORT utilities (DFSORT, SYNCSORT etc.) you can specify the same file in SORTIN and in SORTOUT.

O.

dick scherrer · Posted: Fri Mar 02, 2007 10:56 pm

Hello,

If the sort ever abends, what will you do to re-create the input thas been destroyed?

prav_06 · Posted: Sat Mar 03, 2007 4:35 pm

hallecodec,
U can specify the same dataset name in u r SORTIN and SORTOUT dd st. Yes as Dick mentioned if u r JCL abends u would be loosin all the data, usually this way of sorting is done only when u have a back up of the ds in a GDG base. We have a prog. generating a file with data, the next step of mine is to take back up of the same to GDG , and the third step sorts the file and gives the output on the same file,

Thamilzan.

kregen · New User Joined: 16 Mar 2006 Posts: 21

Hi,
ich do it as follwed

//SORT EXEC=DFSORT
.
.
.
.
.
//SORTIN DD DISP=(OLD,DELETE,KEEP),DSN=GDG.ENTRY(0)
//SORTOUT DD DISP=(NEW,CATLG),DSN=GDG.ENTRY(+1)
.
.
.
THE GDG-Base ist defined with LIMIT 1 SCRATCH.
TSO DEFINE GDG ( NAME(GDG.ENTRY) LIMIT(1)

All followed JOBSTEPS are running with GDG-BASE-Name only

//READIT EXEC=PROG
//INPUT DD DISP=SHR,DSN=GDG.ENTRY
.
.
.
cu
kai

dick scherrer · Posted: Mon Mar 05, 2007 8:39 pm

Hello Kregen,

Neat solution

It solves the problem of still having the input in case of an abend as well as producing the "same" dsn as output. While it does create a "new" file, two files exist only briefly and when the step is done, the downstream processes see only "the file".

hallecodec · Posted: Tue Mar 06, 2007 12:53 pm

hi,

thanks for all of your helpful tips and suggestions...