Regexp/Glob

Renegade498 · 03-29-2005, 05:36 PM

Which method is faster? Regular expression, or glob?

Also, if you are trying to catch a particular phrase, would it be faster for the method to try to catch the EXACT phrase or just a part of it?

Example - If I am trying to catch the phrase:

The quick brown fox jumps over the lazy dog.

Would it be faster for glob/regexp to catch the phrase by using...

"The quick brown fox jumps over the lazy dog." ### - Trying to match the whole sentence

"*brown fox jumps over*" ### - Partial Matching

Thanks in advance.

Proud · 03-30-2005, 09:46 AM

Surely you need to match the whole thing or you dont know it's all there...?

TheLinuxDuck · 03-30-2005, 11:19 AM

Renegade:

A simple perl script will test this for you and tell you the answer. The test script basically loops x number of times, and varies the match so that it only hits every so often, using both 'eq' and a regexp:

Code:

#!/usr/bin/perl

#MAIN
{
  my($loopcount)   = 10000000;
  my($matchevery)  = 50000;
  my($checkphrase) = "The quick brown fox jumps over the lazy dog.";

  my($starttime, $endtime, $lookfor, $matchcount);

  print "Looping $loopcount times, match every $matchevery...\n";

  #  eq test
  #
  $matchcount = 0;
  $starttime  = time;
  print "eq: Started at ", localtime($starttime), "\n";
  for(0..$loopcount) {
    $lookfor = ($_ % $matchevery == 0) ? $checkphrase : "Something different";
    $checkphrase eq $lookfor and $matchcount++;
  }
  $endtime   = time;
  print "eq: Ended at ", localtime($endtime), " with $matchcount matches\n";
  print "eq: Time difference in seconds: ", ($endtime - $starttime), "\n";

  #  regexp test
  #
  $matchcount = 0;
  $starttime  = time;
  print "regexp: Started at ", localtime($starttime), "\n";
  for(0..$loopcount) {
    $lookfor = ($_ % $matchevery == 0) ? "quick brown fox" : "Something different";
    $checkphrase =~ /^.*${lookfor}.*$/ and $matchcount++;
  }
  $endtime   = time;
  print "regexp: Ended at ", localtime($endtime), " with $matchcount matches\n";
  print "regexp: Time difference in seconds: ", ($endtime - $starttime), "\n";

  exit;
}

and when run:

Code:

~/perl> ./eq_or_regexp.pl
Looping 10000000 times, match every 50000...
eq: Started at 1051123021053880
eq: Ended at 2851123021053880 with 201 matches
eq: Time difference in seconds: 18
regexp: Started at 2851123021053880
regexp: Ended at 452123021053880 with 201 matches
regexp: Time difference in seconds: 36

Renegade498 · 03-30-2005, 11:52 AM

Alright! Thanks a lot man. That makes thigns crystal clear.