arunhorne has asked for the wisdom of the Perl Monks concerning the following question:

I want to create file containing a set of substitutions, one per line of the form s/foo/bar/g. I then want to load these substs from the file and apply to an arbitrary set of strings (loaded from a different file) to transform them.

Any ideas :)

Replies are listed 'Best First'.
Re: Regexes loaded from file
by samtregar (Abbot) on May 02, 2002 at 18:05 UTC 
    #!/usr/bin/perl -w
open SUBS, "subs.txt"       or die;
my $sub = eval "sub { " . join('', <SUBS>) . "}";
die "Unable to complile subs.txt : $@" if $@;
close SUBS;

open STRINGS, "strings.txt" or die;
while(<STRINGS>) {
  &$sub;
  print;
}
close STRINGS;

    [download]

    -sam

[reply]
[d/l]
      Now why would this get downvoted? That's a working, efficient, tested solution to the problem posted! What do you want, a dance routine to go along with it?

      -sam

[reply]
        Don't worry Sam... I thought this was a neat piece of code u put and it does exactyl what I asked. Nice one :)
[reply]
Re: Regexes loaded from file
by Popcorn Dave (Abbot) on May 02, 2002 at 17:57 UTC
    Seems easy enought to do. Open the file of regexes, open the file you want to change, then read a line from both files, do your substitution and write it.

    My question is: Why you'd want to do this? Homework assignment?

[reply]
      No not homework - needed for felxible processing of a set of metabolite reaction equations. Its not as simple as you allude becuase you need either to use closures or subroutine refs... I just wondered if anyone knew how. Cheers :)
[reply]
        Sorry, thought it sounded simpler than it was. :)

        However I did do something similar myself not too long ago for a program I wrote. This was for a regular expression tutor I wrote ( which I hope to post to craft as soon as I iron out one blasted bug ) so hopefully this will prove useful to you. 

        =head2 If we've had a regular expression with grouping, we need to sho
+w the
       matches we made there too. Unless we count the number of matchi
+ng
       parens, which could get odd if there was a match looking for '(
+' and ')',
       we run a simple counter and check if the match variable exists.
+ If it
       does, we print it out. If no parenthesis are used, $match_resul
+ts holds
       our match.
=cut
       {eval "\@matches = (\$user_text =~ $regex); \$match_results .= 
+\$&;";
             $match_results .= "\n";
             $match_results .= " Variables:\n"if defined($matches[$i])
+;
             for (my $i = 1; $i <= @matches; $i++) {
                  my $found = $matches[$i - 1];
                  $match_results .= " \$$i:$found \n" if defined($matc
+hes[$i]);
                  }
             } # end eval block

        } # end if substr eq 'm'

        [download]

        This does encompass multiple matches on an m//. The s// and tr// I did like: eval "\$match_results =~ $regex

        Of course I did have the help of some here to pull this stunt off :)

        Good luck!

[reply]
[d/l]
        Some sort of <code>while(<>){push @subs,eval 'sub{local$_=shift;'.$_.';$_}'}<code>?
[reply]

Back to Seekers of Perl Wisdom