Create sequential file only if any rows written into it

satheesh_color · Post by **satheesh_color** » Wed Sep 30, 2009 2:05 am

Hi All,

We have a requirement to Create sequential file only if any rows written into it, otherwise no need to create a sequential file.

please let me know you thoughts on how to acheive this task.

Thanks,
Satheesh

ArndW · Post by **ArndW** » Wed Sep 30, 2009 2:32 am

Instead of complicated workaround, put in an after-job call to SH and execute "find . -name {YourFileName} -size 0 -exec rm -f {} \;" . I am sure that the real UNIX experts will suggest a more efifacious command, but this should work well enough.

Sainath.Srinivasan · Post by **Sainath.Srinivasan** » Wed Sep 30, 2009 2:45 am

or you can create the file with a different name and copy it to your expected output name if the size is greater than 0.

Code: Select all

test -s

will assist you.

ArndW · Post by **ArndW** » Wed Sep 30, 2009 2:46 am

I've only used "test -s" in scripts and don't know how it can be used in a single command line to delete zero length files, which is why I reverted to the time-honored and much distinguished "find" command. Do you have an example of how to use "test" without writing or calling a script?

chulett · Post by **chulett** » Wed Sep 30, 2009 5:21 am

Bottom line, Satheesh, is that your sequential file will always be created regardless of how many records are written to it, including zero. So you will either need some kind of 'work around' to fake this (as posted) or just live with and don't worry about it (which is what practically everyone does).

satheesh_color · Post by **satheesh_color** » Thu Oct 01, 2009 2:00 am

Hi,

Thanks for your feedback on this. We made a decision to overwrite the file everyday irrespective of number of records.

Thanks,
Satheesh

Sreenivasulu · Post by **Sreenivasulu** » Thu Oct 01, 2009 3:28 am

You can use the 'touch' command as well

Regards
Sreeni