Dropping Columns off a sequential file

[wjfitzgerald]

Hi,

I need to drop a number of fields off each record in a file and be left with the new format under the old file name.

eg input file is called ext/file1 and has four columns, after the process drop 2 colums, i still have a file called ext/file1 with just the two remaining fields.

I have no idea how to keep the file name, or if this is even possible.

As usual, any suggestions would be gratefully recieved, thanks in advance.
Regards,

John Fitz

[ArndW]

You need to write the truncated output to a 2nd file, then use an after-job routine to call the UNIX "mv -f /path/ext/file2 /path/ext/file1"

[wjfitzgerald]

thanks for that, i will look into that straight away.
Could i ask one further question, if i am using a file pattern to read in the file, is there any way to pass out the file name to the unix command, or am i really over complicating the issue at this point?

Thanks again,

john fitz

[ray.wurlod]

Each column in the record schema has a Drop On Input property that you can use to specify that that particular column does not get read into the job.

This is only true in the parallel job Sequential File stage.