Appending data to the same Sequential file

[Simba]

My job is creating four individual sequential files of the same structure. I want to append data in to the same seq file, instead of using four files. I have tried ExecSh with touch command with a sample job of two seq files and loading into the same seq file. For first time execution target seq file has only records from source seq file1, for next run target file has data from both source files 1 and 2. Some thing wrong i guess. please send me if any one of you has routine to create/delete a sequential file.

Thanks,

[kcbland]

You cannot have multiple processes writing to the same sequential file, period. There's nothing that can be done about this because it's a block writing issue on all operating systems. You have a few choices: if all streams are in one job use a link collector; if separate jobs use a hash file to collect the data using a unique meaningless key to prevent row collisions; else use four separate sequential files and cat them into one and remove the quarter sized original files.

[ray.wurlod]

Sage advice.

You can cat the remaining three to the first, and make the rm command conditional on success of the cat command, using ordinary shell operators, for example

Code: Select all

cat file2 file3 file4 >> file1 && rm -f file2 file3 file4

The same can be done on Windows, use type instead of cat and del instead of rm.

Code: Select all

type file2 file3 file4 >> file1 && del /Q file2 && del /Q file3 && del /Q file4

(you can use wildcards with del, but not multiple file names; there's also a variant of copy that you could use)