Perl RSS Perl RSS

Remove leading and trailing white space from a string

KevinADC KevinADC is offline Offline Sep 4th, 2006, 3:28 am |
0
This is a routine task best done with two regexp's when using perl.
Show Printable Version Show Printable Version
Quick reply to this message  
Code Snippet
•
•
•
•
Perl Syntax
Toggle Plain Text
  1. my $string = '  Mary had a little lamb.  ';
  2. $string =~ s/^\s+//; #remove leading spaces
  3. $string =~ s/\s+$//; #remove trailing spaces
  4. print $string;
Code Snippet Comments
0
sut sut is offline Offline | Oct 3rd, 2006
•
•
•
•
Or:

$string =~ s/^\s*(\S*(?:\s+\S+)*)\s*$/$1/;
 
0
KevinADC KevinADC is offline Offline | Oct 4th, 2006
•
•
•
•
Yes, but benchmarking the two differnt ways is revealing:


Perl Syntax (Toggle Plain Text) 
  1. my $string = "    Mary had a little lamb.   ";
  2.  
  3. my  $results = timethese(200000, 
  4.         {
  5.             'First' => sub {$string =~ s/^\s*//;$string =~ s/\s*$//;},
  6.             'Second' => sub {$string =~ s/^\s*(\S*(?:\s+\S+)*)\s*$/$1/;}
  7.         },
  8.     );
  9. cmpthese( $results ) ;

results:


Perl Syntax (Toggle Plain Text) 
  1. Benchmark: timing 200000 iterations of First, Second...
  2.      First:  1 wallclock secs ( 2.25 usr +  0.00 sys =  2.25 CPU) @ 88888.89/s (n=200000)
  3.     Second:  3 wallclock secs ( 3.73 usr +  0.00 sys =  3.73 CPU) @ 53619.30/s (n=200000)
  4.           Rate Second  First
  5. Second 53619/s     --   -40%
  6. First  88889/s    66%     --

using two regexps is 66% percent faster (on my computer, results will vary).
 
0
rockslammer rockslammer is offline Offline | May 14th, 2008
•
•
•
•
$string =~ s/\s+$//; #remove trailing spaces
Worked very nicely for me. The expresion solved an LWP issue -- thanks
 
1
foobie foobie is offline Offline | Nov 28th, 2008
•
•
•
•
Actually, $string =~ s/^\s+//;$string =~ s/\s+$//; is faster, matching only when there is one or more whitespace characters.

Your Benchmark is flawed and only strips whitespace on the very first run - $string is global.

Try this benchmark instead:
perl Syntax (Toggle Plain Text) 
  1. #!/usr/bin/perl
  2.  
  3. use warnings;
  4. use strict;
  5.  
  6. use Benchmark qw(cmpthese timethese);
  7.  
  8. sub double_star {
  9.   my $string = shift;
  10.   $string =~ s/^\s*//;
  11.   $string =~ s/\s*$//;
  12.   return $string;
  13. }
  14.  
  15. sub double_plus {
  16.   my $string = shift;
  17.   $string =~ s/^\s+//;
  18.   $string =~ s/\s+$//;
  19.   return $string;
  20. }
  21.  
  22. sub replace {
  23.   my $string = shift;
  24.   $string =~ s/^\s*(\S*(?:\s+\S+)*)\s*$/$1/;
  25.   return $string;
  26. }
  27.  
  28. sub for_star {
  29.   my $string = shift;
  30.   for ($string) { s/^\s+//; s/\s+$//; }
  31.   return $string;
  32. }
  33.  
  34. sub for_plus {
  35.   my $string = shift;
  36.   for ($string) { s/^\s*//; s/\s*$//; }
  37.   return $string;
  38. }
  39.  
  40. sub regex_or {
  41.   my $string = shift;
  42.   $string =~ s/(?:^ +)||(?: +$)//g;
  43.   return $string;
  44. }
  45.  
  46. cmpthese(
  47.   -1,
  48.   {
  49.     'double_star' => q|double_star('    Mary had a little lamb.   ');|,
  50.     'double_plus' => q|double_plus('    Mary had a little lamb.   ');|,
  51.     'replace'     => q|replace(    '    Mary had a little lamb.   ');|,
  52.     'for_star'    => q|for_star(   '    Mary had a little lamb.   ');|,
  53.     'for_plus'    => q|for_plus(   '    Mary had a little lamb.   ');|,
  54.     'regex_or'    => q|regex_or(   '    Mary had a little lamb.   ');|,
  55.   }
  56. );

Results:
Perl Syntax (Toggle Plain Text) 
  1.                 Rate regex_or  replace for_plus double_star for_star double_plus
  2. regex_or     55855/s       --     -47%     -49%        -60%     -73%        -84%
  3. replace     105217/s      88%       --      -5%        -25%     -49%        -70%
  4. for_star    110277/s      97%       5%       --        -22%     -46%        -68%
  5. double_star 140894/s     152%      34%      28%          --     -31%        -59%
  6. for_plus    204799/s     267%      95%      86%         45%       --        -41%
  7. double_plus 345717/s     519%     229%     213%        145%      69%          --
 
0
KevinADC KevinADC is offline Offline | Dec 10th, 2008
•
•
•
•
foobie,

excellent post and a good observation about my flawed test.

Regards,
Kevin
 
0
acca acca is offline Offline | 28 Days Ago
•
•
•
•
Another one:

sub chomp_plus {
my $string=shift;
$string =~ s/^\s+//;
chomp $string;
return $string;
}
Perl Syntax (Toggle Plain Text) 
  1.                 Rate regex_or replace for_plus double_star for_star double_plus chomp_plus
  2. regex_or     98642/s       --    -50%     -55%        -63%     -71%        -81%       -88%
  3. replace     196495/s      99%      --     -11%        -26%     -42%        -63%       -77%
  4. for_plus    220554/s     124%     12%       --        -17%     -35%        -58%       -74%
  5. double_star 265481/s     169%     35%      20%          --     -22%        -50%       -69%
  6. for_star    341333/s     246%     74%      55%         29%       --        -35%       -60%
  7. double_plus 526091/s     433%    168%     139%         98%      54%          --       -38%
  8. chomp_plus  849541/s     761%    332%     285%        220%     149%         61%         --
 
 

Quick Comment
•
•
•
•
•
•
•
•
Message:
•
•
•
•


Similar Threads
Thread Tools Search this Thread



access array asp cgi checkbox hash html http multiplewindows parse parsing perl perlsystemwindows perltk post script search split string table text text-file
About Us | Contact Us | Advertise | DaniWeb | Acceptable Use Policy | RSS Feed

©2003 - 2009 DaniWeb® LLC