Re: [Help-bash] how to grep the last match out from a bunch of files

[lina]

On Wednesday 28,December,2011 02:00 AM, Greg Wooledge wrote:
> > On Wednesday 28,December,2011 01:12 AM, Greg Wooledge wrote:
> > > awk '/^step$/ {getline; step=$0} END {print step}'
> On Wed, Dec 28, 2011 at 01:33:53AM +0800, lina wrote:
> > sorry, too concise for me to be able to put into practice about the
> > bunch of files part.
> > $ more try.sh
> > #!/usr/bin/awk
> >
> > awk 'BEGIN{
> >      /^step$/
> >      }
> >      {
> >      getline;
> >      step=$0
> >      }
> > END{
> >      print step
> > }' $0
> You do not want to use BEGIN here.  You want the /^step$/ pattern to be
> matched against each line of input.
>
> If you want it to be an "awk script", then do this:
>
> #!/usr/bin/awk
> /^step$/ { getline; step=$0 }
> END      { print step }
>
> Both this, and my original, only operate on *one* file at a time.  If
> you want to run it against multiple files, then you either need to
> rewrite it, or just run it from a shell loop:
>
> for f in a_*.txt; do
>    awk '....' "$f"
> done
Thanks, I am not familiar with awk so well.

$ for f in a_*.txt; do echo -n -e "$f\t" ;  grep -A 1 step $f | tail -1 
; done
a_1.txt    3
a_2.txt    4

$ for f in a_*.txt ; do echo -n -e "$f\t" ; awk '{/^step&/ getline; 
print $0}' $f | tail -1; done
a_1.txt    3
a_2.txt    4

Come up above 2 with your hints.

Best regards,