Merge 5 files,create one file with count in HDR & TRL

puneetvirmani143 · New User Joined: 24 Jul 2007 Posts: 55 Location: noida

Hi,

I will be getting 5 files from offhost system (which are junks of same file) to be combined together on mainframe. All have LRECL=4000, RECFM=FB and have header & trailer

1. Header & Trailer are in following format:
a) HEADER<<FILENAME>><<COUNT of detail records>>
b) TRAILER<<COUNT of detail records >>

Note: Count of detail records appearing header and trailer are equal in each respective file. FILENAME will be same in all 5 files.

2) key field in detail record starts at Position 1 and length 11 bytes.

Output Expected.

1. Combine all 5 files so that details records appear in sorted order of Key field and duplicates are rejected on the basis of Key field.

2. Only 1 header and 1 trailer should appear in the combined file with count appearing in Header & trailer to match with the count of detail records in Output combined file.

3. while creating header for output file, Pos1 - Pos 20 of header (all byte position before the start of count) should be used. These Byet pos1-Pos20 will be same in all 5 files.

Please suggest.

puneetvirmani143 · New User Joined: 24 Jul 2007 Posts: 55 Location: noida

Can anyone please help with this requirement ? Thanks.

enrico-sorichetti · Posted: Sat Mar 02, 2013 1:52 pm

it' s saturday, nobody works...
so You have no reason to solicit for an answer after 3 hours of the initial post.

You should remember that replying is ...
on voluntary base
free of charge
interest of the topic

in less PC words... when and if we feel like

if You have time constraints, a forum is not the best place to ask

problems like this one have been discussed soo many times that have become pretty boring

did You search the forum before soliciting for a reply ???

some of the topics having hints on the solution were even started by YOU !

Bill Woodger · Posted: Sat Mar 02, 2013 2:58 pm

puneetvirmani143,

As enrico has pointed out, you should already have sufficient knowledge from last December about how to do this, including the "stupidity" of having the record-count on the header.

With the record-count on the header, you know for certain that the header has been added after the creation of the data-records. So, until the point that the header is added, the file is "unprotected" by it, and lots of things can happen to mess it up before something comes along and stamps it "This file is fine". Nuts.