Splitting input file into multiple output files

PORYES · New User Joined: 04 Apr 2007 Posts: 38 Location: Chennai

Hello,

I want to split my input file into multiple o/p files.
Input file has thousands of records.

I want to split in to 15000 record in to one O/P1 and other 15000 to O/P2 and so on.. But before splitting the file I want to check the last record is NOT "A" at first position.

I/P contains

A 2013
1234
B 21222

-
0 11221 TOTAL AMPUNT 2110.22

A12 2013
2235
B 41222
0 1111TOTAL AMPUNT 543.22
A02 2013

My outout should be

O/P1

A 2013
1234
B 21222

-
0 11221 TOTAL AMPUNT 2110.22

O/P2

A12 2013
2235
B 41222
0 1111TOTAL AMPUNT 543.22

O/P3

A02 2013

record format - FBA
lrecl-133

Can anyone suggest me how to acheive this using sort, icetool

thanks!

Bill Woodger · Posted: Sun Oct 27, 2013 4:57 am

Have a look here.

PORYES · New User Joined: 04 Apr 2007 Posts: 38 Location: Chennai

Hi Bill,

I'm not not clear that will satisfy my requirement.
My input has thousands of records...

For eg:
Input
A 2013
1234
B 21222
blanks
-
A 2013
1234
B 21222
o
blanks
A 2013
1234
B 21222

And want to split each out put should have 15000, while splitting It has to split the before record "A". While checking the condition the output doesn't have exactly 15000. but it should close to 15000, ie 14990 to 15100. The last output file should contain the rest of the records.. But the first record should start with "A"

Hope its clear.. Any help is appreciated..

Bill Woodger · Posted: Mon Oct 28, 2013 1:06 pm

Well, how about becoming clear whether it will satisfy your requirement? What is it specifically which makes you think it won't?

"Split a file but keeping groups together" seems to be what you want.

PORYES · New User Joined: 04 Apr 2007 Posts: 38 Location: Chennai

I want the first 15000 records in OUT1 but the last record shouldn't be "A" record. Next 15000 in OUT2 but the first record should be "A" record.

PORYES · New User Joined: 04 Apr 2007 Posts: 38 Location: Chennai

Ran using the sort card and I'm getting the below error.

OPTION COPY
INREC IFTHEN=(WHEN=INIT,OVERLAY=(134:SEQNUM,8,ZD)),
IFTHEN=(WHEN=GROUP,KEYBEGIN=(1,5),PUSH=(144:139,4))
*
OUTFIL FILES=01,INCLUDE=(144,4,ZD,EQ,00),BUILD=(1,133)
OUTFIL FILES=02,INCLUDE=(144,4,ZD,EQ,01),BUILD=(1,133)
OUTFIL FILES=OV,SAVE,BUILD=(1,133)
WER268A INREC STATEMENT : SYNTAX ERROR
WER211B SYNCSMF CALLED BY SYNCSORT; RC=0000
WER449I SYNCSORT GLOBAL DSM SUBSYSTEM ACTIVE

Any help

Akatsukami · Posted: Tue Oct 29, 2013 12:23 am

Please post the entire body of SYSOUT messages in Code tags.

PORYES · New User Joined: 04 Apr 2007 Posts: 38 Location: Chennai

Here is the entire jcl

Bill Woodger · Posted: Tue Oct 29, 2013 6:19 am

I can't see anything wrong with either of the INCLUDEs, but your BEGIN is malformed. You have taken a DFSORT KEYBEGIN and chopped off the word KEY, I think.

SyncSort doesn't have KEYBEGIN. If you need to emulate it, look at SEQNUM with RESTART=

gcicchet · Posted: Tue Oct 29, 2013 6:34 am

Hi,

I think there is another error you are not showing.

Gerry

PORYES · New User Joined: 04 Apr 2007 Posts: 38 Location: Chennai

Is there a way I can extract multiple blocks,

ie 1 st record to 15000, and check the last record before write?

Bill Woodger · Posted: Tue Oct 29, 2013 2:56 pm

Yes. You've made a start.

You make a sequence number for each record.

You make a GROUP using BEGIN= with the value "A". For this you PUSH the sequence number on to each record of the group.

You use the PUSHed sequence number in the OUTFILs for selection.

If the 15,000th is an A, all records of that group will have the PUSHed sequence number as 15000.