Regex?

[safra]

Hi,

The options in a combobox all have something added to it which takes 5 characters.

What would be the quickest way to cut of these 4 variables in the Perl script?

example:

$var = "something (ma)";

should become:

$var = "something";

If I try long enough I might be able to get there with substring but I guess this can be done a lot quicker?

regards,

Ron

 

[luciddream]

well, you could you reverse($var);

then take the substring then reverse it again.

or you could use

$var =~ s/^(.+?).{4}$/$1/g;

or something along those lines. i didn't test it, but, it should work. adam@aauser.com

 

[hmerrill]

A slightly different twist would be to do a substring this way:

my $a = "abcdefghit(kl)";
my $b = substr($a,0,length($a)-4);
print &quot;\$a = <$a>\n&quot;;
print &quot;\$b = <$b>\n&quot;;

Here's the output:
$a = <abcdefghit(kl)>
$b = <abcdefghit>

HTH.
Hardy Merrill
Mission Critical Linux, Inc.

http://www.missioncriticallinux.com

 

[luciddream]

hmm... what the hell was i thinking? reversing does no good in this situation. adam@aauser.com

 

[safra]

Thanks quys,

Luciddream I tried:

$var =~ s/^(.+?).{4}$/$1/g;

But it didn't seem to work?

I will try your suggestion hmerrill.

btw. yesterday I read something in the perl docs that regex are relatively slowing down the process more than something like hmerrill suggested. Is this true, I thought regex were quite handy and &quot;fast&quot; pieces of code?

Ron,

 

[hmerrill]

I'd be interested in reading that if you can post what you find that says regex's are slow. I think that some complex regex's may be slower than simple regex's, but I have trouble believing that in general regex's are slow. In fact, I just tested it - here's my test script - and luciddream's $var =~ s/^(.+?).{4}$/$1/g example did work for me - I just found this &quot;times&quot; function today by accident in the Programming Perl O'Reilly book on p.234:

#!/usr/bin/perl -w

use strict;

my ($user, $system, $cuser, $csystem) = times;
my $a = &quot;abcdefghit(kl)&quot;;
my $b = $a;

my $start = (times)[0];

for (my $i = 0; $i < 100000; $i++) {
$b = substr($a,0,length($a)-4);
}
#for (my $i = 0; $i < 100000; $i++) {
# $b =~ s/^(.+).{4}$/$1/g;
#}

my $end = (times)[0];

printf &quot;that took %.2f CPU seconds\n&quot;, $end - $start;

---------------------------------------

Run it one way, then switch which for-loop you comment out, and run it again. Turns out that luciddream's regular expression takes 0.14 cpu seconds, and my substring takes 0.27 cpu seconds - the substring takes about twice as long as the regular expression. That is for 100,000 iterations of a for loop. Pretty interesting - regular expressions are pretty fast

HTH.
Hardy Merrill
Mission Critical Linux, Inc.

http://www.missioncriticallinux.com

 

[luciddream]

i've never heard of perl regex's being slow... but, i do know that javascript regex's (modeled after perl's) are practically the slowest processing you can do. adam@aauser.com

 

[safra]

I read something in the perl docs on

http://www.perl.com.

I probably misinterpreted it. I'm only a beginner and, as I said, to me regex appeared to be powerful, that was why I was so amazed to read this.

Ron