in reply to Transpose some spaces to commas in a string

Yes, a regex could do that, but you probably don't want to use a regex for this. Text::CSV is a better solution overall. Your main obstacle, however, is figuring out how to identify the persons's name, so you don't break the fields up incorrectly.

Cheers,
Ovid

Join the Perlmonks Setiathome Group or just go the the link and check out our stats.

  • Comment on (Ovid) Re: Transpose some spaces to commas in a string

Replies are listed 'Best First'.
RE: (Ovid) Re: Transpose some spaces to commas in a string
by cosmicv (Initiate) on Sep 29, 2000 at 20:55 UTC
    Thats a problem, some user names are Bob Q Smith, Jane Doe, Ralph Dilbet Johnson, only was I see to do it is to detect the 4 "0000" in that serial number then jump back past the previous word and replace that space with a comma. I have no clue on how to do that. Any tips?
[reply]
      Offhand, I'd say try the FAQ about putting commas into a number. If the variable mess is at the start of your string, reverse it.
      Bob Q Smith bsmith 00001234567 5/1/00 12:00:00
      becomes
      00:00:21 00/1/5 76543210000 htimsb htimS Q boB

      And a fairly simple regex can insert your commas. Re-reverse your result, and there you go.

      Update: here's my attempt that appears to do what you want: 

      #!/usr/bin/perl -w
use strict;

my $test="Bob Q Smith bsmith 00001234567 5/1/00 12:00:00";
$test = reverse $test;
#documenting the simple but nasty looking regex is left as an exercise
+ for the reader
$test=~s/([:\d]+)\s([\d\/]+)\s([\d]+)\s([\w]+)\s(.*)/$1,$2,$3,$4,$5/ o
+r die "Pattern didn't match: $test";
$test = reverse $test;

print $test;

      [download]
[reply]
[d/l]
      Untested: 
      my $testvar = "Bob Q Smith bsmith 00001234567 5/1/00 12:00:00";

$testvar =~ s/
              \s           # a whitespace character (space, tab, etc)
              (            # Capture to $1
                \w+        #    one or more word characters (e.g. bsmi
+th)
              )
              \s           # a whitespace character
              (            # Capture to $2
                \d+        #    one or more digits
              )
              \s           # a whitespace character
            /,$1,$2,/x;

      [download]
      If your data is relatively clean, I think that should work. It relies on you knowing that the first digits you encounter are going to be your serial number and immediately preceding them are some letters. From there, it's pretty straightforward. There are many different ways to write this regex and this may not be the best, but it's fairly clear.

      Cheers,
      Ovid

      Join the Perlmonks Setiathome Group or just go the the link and check out our stats.

[reply]
[d/l]
        Nice one as that works on names with zero or more middle initials, and isn't bothered by capitalized usernames either (as my quick attempt was).
[reply]

      Than the first thing you need to do is write something to clean up the data. Maybe throw the cleaned data into and array, run the regex on that and then output those results to a file...

      Maybe if you could change the input so it comes as something like:

      Bob Q Smith 00001234567 bsmith 5/1/00 12:00:00

      It would make it much easier then to use regexes to catch the spaces that fall between number and letter fields. I think this goes under the "it would be easier to control the input if you can" department.

      -- I'm a solipsist, and so is everyone else. (think about it)

[reply]

In Section Seekers of Perl Wisdom