Pattern match assistance

[tonykent]

I've got a data file containing lines in 2 basic formats e.g.

1 f none /usr/stdapp/cnvcmd/cnvcmd.ast 0666 root other 22681
and
0003d000gh0root/usr/.install/stdapp/defapp/cat008/enccat.ast

I want a pattern match will will pull out the file name with its path from both types of rows. I can get it to work with one or the other but not both! Can you do better? Thanks.

 

[justice41]

For the two lines you posted this works:

m!(/.*?)[^/\w.]!; # file captured into $1

You'll have to be more specific about the format if that doesn't work in all cases.

jaa

 

[mikevh]

/(\/\S+)/ && print "$1\n";

This prints both file names.
The pattern just means "a slash followed by one or more
non-whitespace characters."

 

[dummy33]

this could work too.. in case there's a space in the filename

m ~ ( /.*/        # everything between first and last slash
          \w+     # followed by a word
              \.  # a dot
          \w+)    # another word
      ~x;

-or-

m~( /.*/\w+\.\w+)~x;

 

[dummy33]

Actually, I was thinking if there was a space in the path.. mine won't work if there's a space in the filename.

 

[icrf]

can you guarantee one of two things:
a) the filename ends in '.ast'
b) the filename ends with a three character extention, ie /\.\w{3}/ ----------------------------------------------------------------------------------
...but I'm just a C man trying to see the light

 

[tonykent]

Thanks for all the responses guys. I'll try them out when I get into work tomorrow.

The filename extensions vary and are not all 3 letters long.

 

[tonykent]

Thanks very much to you all. I went with dummy33's solution in the end. The data file was somewhat more complicated than I went into but this solution coped with it all successfully.