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

Hi monks!
I have some bacterial organisms names that I want to look up into a big datafile.
My problem is that i have cases like the following:
[NAME I HAVE]: Acidovorax JS42
[NAME IN THE DATASET FILE] Acidovorax sp. JS42

[download]
So, is there a way that I can use a reg exp to match this? I tried:
$my_name="Acidovorax JS42";
$db_name="Acidovorax sp. JS42";

if($total=~/$find/)
{
    print "OK\n";
}

[download]
and it did not work... So, is there not any way to match a string of some words that is contained in a bigger string but the words are not in order?
Thanks!

Replies are listed 'Best First'.
Re: Reg exps?
by Ratazong (Monsignor) on Jan 14, 2010 at 15:09 UTC

    So, is there not any way to match a string of some words that is contained in a bigger string but the words are not in order?

    1. split your name into various strings
    2. check if each of these strings are in your DB-string (e.g. using substr() instead of an regex)

    HTH, Rata
[reply]
[d/l]
[select]
      Just a thought: Maybe you could (after splitting your string in substrings) do a match if you counted (set $counter for each match) e.g. 2 strings (out of 3 or whatever you think is ok) are in the "complete string" or something like that? After testing that to a larger amount of data u maybe adjust this till it fits ...


      hth
      MH
[reply]
Re: Reg exps?
by eff_i_g (Curate) on Jan 14, 2010 at 15:58 UTC
    Is it safe to assume that the beginning (Acidovorax) and end (JS42) must match, but the middle (sp.) doesn't have to? If so, you can split on a space and build a regex to enact this.
[reply]
      The "my_name" string is definitely included in the "db_name" string, but we can't know beforehand if the "db_name" string contains other characters or words as well. But we sure know that it contains all words from "my_name" string.
[reply]

        I hope I did understand your problem. Here is my sample code. Kindly check it if the flow and output fits to the solution of your problem. If it didn't, I hope it can give you some clues. Just like Javafan said check each word of "my_name" if they all exist in one of the string of database "db_name". 

        #!/usr/bin/perl -w

@my_name=("Acidovorax","JS42");    #name to be searched...
                                   #...split by term/words
@db_name=("Acidovorax sp. JS42",    #data base of names in array
          "JS42 Acidovorax sp.",    
          "JS42 sp. Acidovorax",
          "JS53 sp. Acidovorax",
          "JS42 sp. Axidovorax",
          "JS42Acidovorax sp. "
          );

my $ctr;
#----compare strings in $db_name 1 at a time
foreach my $db_each (@db_name){            
   
   #----search each term/word of @my_name in $db_each
   foreach($ctr=0; $ctr<=$#my_name; $ctr++)   
   {
          #----if a term/word not found break the loop
       last if($db_each !~ /\b$my_name[$ctr]\b/i);  
   }                                                
   #----this will be true if inner foreach didn't break
   if($ctr==$#my_name+1)                            
   {
       print "$db_each\n";       #print the matched name
   }
}
<>;

        [download]

        I have mention the "term/word" in the comment. It might be that two words in your "my_name" is considered as one term or two or more words separated by space.

        The code above does not support if a word should exist 2 or more times in any order. example: my_name = "high class bacteria f7-52 high fever". Just mentioning this condition for more flexibility to your program. It's still up to you.

[reply]
[d/l]
        This is more or less redundant after toastbread's response and quite possibly naive but I've done it now and it seems to do what the OP wanted: 
        use strict;
use warnings;

my $my_name="Acidovorax JS42";
my $db_name="Acidovorax sp. JS42";

my @names = split " ", $my_name;
my $numel = @names;  #number of words to match

my $count;
foreach (@names) {
  if ($db_name=~/^|\s\Q$_\E$|\s/) {
     $count++;  #count matches
  }
}

print "MATCH" if $count==$numel;  #report match if all words match

        [download]
        --->Updated to fix regex - thx JavaFan
[reply]
[d/l]
Re: Reg exps?
by JavaFan (Canon) on Jan 14, 2010 at 15:29 UTC
    I don't understand your code fragment. It doesn't say what $total and $find contain, and it doesn't use $my_name or $db_name.

    So, is there not any way to match a string of some words that is contained in a bigger string but the words are not in order?
    Sure there is, but that's a way of matching that has nothing to do with regexpes. A way of doing what you want is to split both strings into "words" (for whatever definition of "word" is appropriate for your problem domain - there's no universal way of splitting a string into words), then look for each of the words in your smaller string if it's in the larger string. There's a match if and only if all the words in the smaller string are found in the larger string. You should also determine when there's a match if the smaller strings contain a word more than once - may they be matched with the same word in the larger string, or must they match on different positions?
[reply]
      Oh, sorry, typo mistake: 
      
$my_name="Acidovorax JS42";
$db_name="Acidovorax sp. JS42";

if($my_name=~/$db_name/)
{
    print "OK\n";
}

      [download]
[reply]
[d/l]

Back to Seekers of Perl Wisdom