Shell Programming and Scripting

This log file is wacky.

the syntax puts this in the Installation line:
Installation PATCH75682.91 of PATCH75681 complete
Installation PATCH76537.91 of PATCH76537 complete
Installation PATCH92217.91 of PATCH92217 complete

So I'm looking for a sed 's///' to remove the first PATCHxxxx occurance,
but if I try

Code:

sed 's/PATCH[0-9]//' infile > outfile

I still have xxxx.xx in the line. Do I need [0-9] for each digit?

See if this works for you:

Code:

sed 's/PATCH.* of//' input_file

Almost, it's cutting out the 'of' and some lines ref a preview mode before the 'of'.

the above suggestion work perfecly:

issue with your command was a missing * and . in search pattern

Your code should look like

What about this:

Code:

sed 's/ PATCH.* of/ of/1' input_file

dba_frog, next time, specify the desired output.

My apologies, I didn't see the preview in there until reviewing the output

but, this worked

Code:

sed 's/PATCH[0-9.]*//' infile > outfile

and i was very close, just needed a little tweaking.

Code:

awk '/PATCH/{$2=""}1' infile