Condition to be included in Spilt1r

Anitha Sekar · New User Joined: 17 Jan 2009 Posts: 5 Location: Chennai

I have to split the input file in to 10 output files based on the number of records present in the input file. please provide me the jcl

Pandora-Box · Posted: Tue Jul 02, 2013 12:55 pm

Anitha Sekar · New User Joined: 17 Jan 2009 Posts: 5 Location: Chennai

for example, if i am having 40000 records. first 4000 records should go into 1st file, second 4000 into 2nd file,etc

If i am having 50000 records,first 5000 records should go into 1st file.

I searched it. But i didn't find anything.As it is urgent, I have posted it here.

expat · Posted: Tue Jul 02, 2013 1:49 pm

Unfortunately the only URGENT is yours

It would also help if you actually answered the question regarding the sort product being used that Pandora has asked.

I have just done a quick search on the forum for your question and had 150+ hits.

Anitha Sekar · New User Joined: 17 Jan 2009 Posts: 5 Location: Chennai

sort,sychsort,iceman,icetool are the sort product being used

Pandora-Box · Posted: Tue Jul 02, 2013 2:08 pm

Anitha,

I guess you use SYNCSORT

Try this

Arun Raj · Posted: Tue Jul 02, 2013 3:31 pm

Anitha Sekar,

With SPLIT, rec1 from input goes to output1, rec2 goes to output2 and so on. Is that what you are expecting?

Anitha Sekar · New User Joined: 17 Jan 2009 Posts: 5 Location: Chennai

Hi Arun,

I have to find the count the number of records present in input file and divide by 10 to put it into 10 output files.

Please help me on this..

expat · Posted: Tue Jul 02, 2013 5:37 pm

Is there any perticular reason why the solution given by Pandora is not acceptable.

That way you do not need to know anything about record counts at all.

Perhaps if you had given a wider view of the problem, an explanation why they must be in consecutive chunks, then that information would help people help you without wasting time and effort.

enrico-sorichetti · Posted: Tue Jul 02, 2013 5:46 pm

Pandora' s solution will split data with a <rotating> scheme

and that might not be what the TS wants

but if the TS does not tell we will never know

Pandora-Box · Posted: Tue Jul 02, 2013 6:44 pm

My solution will not work as per the requirement but TS needs to implement what Arun has said

My solution will just be an example for the Urgent TS

Bill Woodger · Posted: Wed Jul 03, 2013 1:22 am

Have a look at this one from the Beginner's site. It was there this morning, you could have had the thing finished before you even posted if it was that urgent.

I haven't checked, but I'm sure there are other examples which are easy to find and apply.

dick scherrer · Posted: Wed Jul 03, 2013 6:08 am

Hello,

The way to do this using the least machine resources is to have the process that created the whole file create another dataset that has the record count.

Then read the record cont file and generate the OUTFIL statement.

Run the split.