Split a file using DFSORT

[ashykhanna]

Hi,
I have a huge file and i want that to be split into two.
I want the first file to contain all the records till a given condition. Suppose i have

100name Lowerlevel
101a1111programer
102a2222engineer
103a2323analyst
104a3434managers
105a2343smanager
106a4322Dmanager
i want the first file to have all records till i find a managers record and the second file should contain the rest. How can i achieve this using dfsort (no icetool).

file1

100name Lowerlevel
101a1111programer
102a2222engineer
103a2323analyst

file2

104a3434managers
105a2343smanager
106a4322Dmanager


Thanks,
Ashok

 

[COBOLKenny]

I don't know that you can do this with one execution of SORT since there is only one output DD in the JCL (SORTOUT). But you could get it done with 2 SORTs. In the first, you could use the OMIT control statement to omit all records that have "manager" in a certain column. In the second execution you could use the INCLUDE control statement to include only the records that have "manager" in a certain column.

 

[3gm]

Kenny -

I don't think that's what the poster had in mind. I think is was basically "copy to file 1 until this value shows up" and copy all following records to file 2. If one knew the record counter of the split point, one could do this easily.

I suppose there's a way to get the record count with SORT using the COUNT field and INCLUDE. One could then process that record/records with REXX to create a MERGE FIELDS=COPY job that produced two outputs; one using ENDREC and the other using STARTREC. Interesting exercise; ugly, but interesting.

Regards.

Glenn

 

[WMK]

It sounds to me that you are looking for the

OUTFILE

operation. See:

http://publibfp.boulder.ibm.com/cgi-bin/bookmgr/BOOKS/ice1ca00/3.13

which says (among other things),

"Creation of multiple output data sets containing different ranges or subsets of records from a single pass over one or more input data sets. In addition, records that are not selected for any subset can be saved in a separate output data set."

The very first example at:

http://publibfp.boulder.ibm.com/cgi-bin/bookmgr/BOOKS/ice1ca00/3.13.2.1

seems to me to be what you are asking for.

Bill Klein

 

[WMK]

P.S. to my last post.

Of course, cince you asked this in a COBOL forum, it would be a "piece of cake" to do this in a COBOL program with an internal SORT - with an Output Procedure <G>.

Bill Klein

 

[ashykhanna]

Let me be more clear..

100name Lowerlevel
101a1111programer
102a2222engineer
103a2323analyst
104a3434managers
105a2343fdfdf
106a4322gfgfhh

File 1

100name Lowerlevel
101a1111programer
102a2222engineer
103a2323analyst

File 2

104a3434managers
105a2343fdfdf
106a4322gfgfhh

WMK,
I guess the link you sent is not the solution.

can anyone explain me with the code (only JCL - DFSORT, No cobol pls)

Thanks in advance.
Ashok

 

[WMK]

Look at the example at:

http://publibfp.boulder.ibm.com/cgi-bin/bookmgr/BOOKS/ice1ca00/3.13.2.7

and tell me how that doesn't do what you are asking for?

(If you need the JCL and how to modify the code exactly, then I don't think this is a task YOU should be doing - without some help from within your shop.)

Bill Klein

 

[3gm]

Bill -

As I pointed out in my prior post, one has to know a priori what the record number is for the split point. I think the poster wants sort to figure that out somehow.

Glenn

 

[jmanj]

So you want file 1 to contain from 1st record up to the manager record and all the rest goes to file2? If this is the case use a switch and read the file sequentially. Sample below in COBOL.

Move 0 to Switch1.

PERFORM until (end-of-file)
read TheFile.

if Record contain "Manager"
move 1 to Switch1


if Switch1 = 0
add record to File1
else
add record to File2
end-if

end-perform.

 

[WMK]

Did you look at the example with the INCLUDE statement? That does exactly what you are asking for. That supports syntax like

INCLUDE__COND=__ _(logical expression)_

See:

http://publibfp.boulder.ibm.com/cgi-bin/bookmgr/BOOKS/ice1ca00/3.7

Bill Klein