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

I have file that has multiple mixed cased words. I was woundering if there is anyway to do a mass tr/// on any word no matter what the letters are.
Example:
PROjECT_RIgHTS, APPLICATION_RIgHTs
How could I change each set without doing 
$_ =~ tr/PROjECT_RIgHTS/PROJECT_RIGHTS/;
$_ =~ tr/APPLICATION_RIgHTs/APPLICATION_RIGHTS/;<br>

[download]
Or, is this the only way.
Thanks
Bobby

Replies are listed 'Best First'.
Re: tr/<mixed string>/<UPPER STRING>/
by Helter (Chaplain) on Sep 26, 2002 at 14:20 UTC
    Or of course there is:

    uc();
    or.
    "\U$_"
[reply]
[d/l]
[select]
      Yes, and in fact uc is a more robust solution than tr/a-z/A-Z/ because it respects the current locale and Unicode uppercase mappings. ++Helter

      -- Mike

      --
      just,my${.02}

[reply]
[d/l]
[select]
        I'll second that thumbs-down for using tr. The instant you throw in accented words, you're going to cause mayhem. "RESUMé" being a simple example.

        How about this?
        % perl -pi -e '$_=uc' file1 file2 ...
        Careful where you aim it, because it modifies the files in place. Remember to make backups first if you're wary.
[reply]
[d/l]

        In order to make \U work with accented chars I had to start my program like this:

        $loc = 'pt_PT.ISO_8859-1';  # my locale
$ENV{'LANG'} = $loc;
$ENV{'LC_ALL'} = $loc;
$ENV{'LOCPATH'} = '/usr/share/locale';
use locale;
require POSIX;
POSIX::setlocale(&POSIX::LC_ALL, $loc);

        [download]

        Is there a shorter way?

[reply]
[d/l]
Re: tr/<mixed string>/<UPPER STRING>/
by rob_au (Abbot) on Sep 26, 2002 at 14:17 UTC
    How about simply ...

    tr/a-z/A-Z/

    The transliteration operator (tr) is designed to work with lists of characters, replacing the character in the search list with the corresponding character in the replacement list. Additionally a character range can be specified with a hyphen, such that tr/A-E/0-5/ is equivalent to tr/ABCDE/01234/ - See perlop for details.

     

[reply]
[d/l]
[select]
      Thanks, that worked!
      I knew that I should use tr///, but the last time I used
      tr/a-z/A-Z/ it didn't work the way it was suppost to. This time it did.
      Thanks again
      Bobby
[reply]
      Ah, the beauty of perl.
[reply]
Re: tr/<mixed string>/<UPPER STRING>/
by zigdon (Deacon) on Sep 26, 2002 at 14:24 UTC
    you don't have to use tr for this - just use the built in uc function: 
    $_ = uc;

    [download]
    or, if you're using the var in a string, you can use the \U: 
    $_ = "\U$_";

    [download]

    -- Dan

[reply]
[d/l]
[select]
Re: tr/<mixed string>/<UPPER STRING>/
by virtualsue (Vicar) on Sep 26, 2002 at 14:32 UTC
    Here is a filter that will do the job from the command line: 
    perl -pi.bak -e '$_ = uc' your_files

    [download]
[reply]
[d/l]
Re: tr/<mixed string>/<UPPER STRING>/
by Preceptor (Deacon) on Sep 26, 2002 at 14:18 UTC
    cat $file | tr [:upper:] [:lower:] | sed 's/<pattern1>/<pattern2>/ Although I suspect what you want to be doing is something like,
    $_ =~ s/PROJECT_RIGHTS/PROJECT_RIGHTS/gi
    (the 'i' says ignore case)
    --
    It's not pessimism if there is a worse option, it's not paranoia when they are and it's not cynicism when you're right.
[reply]
[d/l]
[select]

Back to Seekers of Perl Wisdom