IBM Mainframe Forum

Hi BillyBoyo,

Thanks for your direction.

Please help me in understanding the below statement in your suggestion.

"WHEN=GROUP on the second field for the indicator which has been PUSHed and not TOTAL, to PUSH the value to the selection marker."

Thanks,
Amar

Hi guys,

I am able to achieve my results using the following three SORT cards.

//SYSIN DD *
SORT FIELDS=COPY
INREC IFTHEN=(WHEN=GROUP,END=(5,5,CH,EQ,C'TOTAL'),
PUSH=(134:ID=4))
OUTFIL FNAMES=OUT1,
INCLUDE=(134,4,ZD,EQ,1),BUILD=(1,133)
OUTFIL FNAMES=OUT2,
INCLUDE=(134,4,ZD,GT,1),BUILD=(1,137)
/*

//SYSIN DD *
SORT FIELDS=(134,4,CH,D)
OUTREC IFTHEN=(WHEN=GROUP,END=(5,5,CH,EQ,C'TOTAL'),
PUSH=(138:ID=4))
OUTFIL FNAMES=OUT1,
INCLUDE=(138,4,ZD,EQ,1),BUILD=(1,133)
OUTFIL FNAMES=OUT2,
INCLUDE=(138,4,ZD,GT,1),BUILD=(1,141)
/*

//SYSIN DD *
SORT FIELDS=(138,4,CH,D)
OUTFIL FNAMES=OUT1,BUILD=(1,133)
/*

I really appreciate your help.

Thanks,
Amar

Run this,



And put the SORTOUT to SYSOUT.

Look at your SORTOUT for the step.

You should see four groups of data.

If you use OUTFIL INCLUDE for ID (134,4,CH,EQ,C´0001´) if you want to keep the ID four long (you only need one long, from what you said).

A second OUTFIL INCLUDE for ID (134,4,SS,EQ,C´0001,0002´) (if you don´t know what the SS is doing, look it up in the manual).

A third OUTFIL with SAVE.

Amar,

you are complicating a simple request. You don't need 3 passes of data as well as a sort operation. use the following DFSORT JCL which will give you the desired results in just 1 pass



If you're not familiar with DFSORT and DFSORT's ICETOOL, I'd suggest reading through "z/OS DFSORT: Getting Started". It's an excellent tutorial, with lots of examples, that will show you how to use DFSORT, DFSORT's ICETOOL and DFSORT Symbols. You can access it online, along with all of the other DFSORT books, from:

http://www.ibm.com/support/docview.wss? ... g3T7000080

Kolusu that' s almost the same solution I proposed, but the TS changed the requirement on the fly...
but the TS rejected it because He has duplicate groups.

enrico-sorichetti wrote:Kolusu that' s almost the same solution I proposed, but the TS changed the requirement on the fly...
but the TS rejected it because He has duplicate groups.

Enrico,

I assumed OP wasn't able to map your solution to 133 bytes from 80 bytes. My bad. If the data is indeed like you shown OP wouldn't get the desired results unless he had EQUALS as default. Even if he had the solution assumes there are only 2 duplicate groups. What happens if he had more than 2 groups? How does he distinguish duplicate groups? REC2 and REC2a group? Is it by the same total for each group or the detailed records have the same indicator?

Hi Kolusu,

In my requirement there are only three groups. Out of which group-1 and group-3 occurs only single time. and Group-2 occurs multiple times depending upon number of records that are being generated. Hence I used three sort cards to achieve my results.

Thanks,
Amar

It will be possible with JOINKEYS, without sorting the data.

Can you confirm genuine sample input data which matches your requirement, and the sample output for that?

Hi Billy,

Please find the sample input file below.



Ouput:

File1:



File2:



File3:


Please note that number of records for Total-2 may vary each time the input file is generated hence total occurrence of Total-2 is not constant. In the above example I gave the number of occurrences as 2, some times it may be 6 or 1 or 9 etc,. Total-1 and Total-3 always occur single time.

Thanks,
Amar

Code'd

OK, untested.

For your first file, in JNF1CNTL, append a "1". This is just to ensure that a key with a constant value is available for the whole file. It is not "relevant" because no matches will occur. You tell the JOINKEYS that this is SORTED.

For your second file, which you will allow JOINKEYS to sort for you, INCLUDE the TOTAL-1 and TOTAL-2 records. Use SUM with a set-up value of '1' to tell the range of the ID for the TOTAL-2 records later (which is why we include TOTAL-1 in the SUM). The "key" is set-up with a value of '0', to arrive before all the '1's of the data.

In the JOIN, ONLY have unmatched records (no matches will be possible, but cleaner to be specific).

In the main-task have an additional GROUP for the unmatched record from file 2, which is our SUMmed total. PUSH this so that it is available on all the records.

The first OUTFIL remains as was shown earlier.

The second OUTFIL has an AND added, so that the full range of IDs can be included for the variable number of TOTAL-2 groups.

The third OUTFIL specifies SAVE, so that anything not written to the first or second OUTFIL will end up here.