RE: remove blanks
On Sep 29, Hanson, Rob said:
>I ran some benchmarks.
>
>The two-liner outperformed the one-liners by a 10 to 1 ratio. Code and
>results below.
>
>Benchmark: timing 10 iterations of OneLine, OneLine2, TwoLines...
> OneLine: 41 wallclock secs (39.30 usr + 0.00 sys = 39.30 CPU) @ 2544.79/s
> OneLine2: 34 wallclock secs (32.58 usr + 0.00 sys = 32.58 CPU) @
>3069.56/s
> TwoLines: 3 wallclock secs ( 2.58 usr + 0.00 sys = 2.58 CPU) @
>38789.76/s
You'll get a better representation with the following benchmark:
#!/usr/bin/perl
use Benchmark 'cmpthese';
my $s = " " . join(" ", ('foo') x 100) . " ";
cmpthese(-5, {
capture => \&capture,
alt => \&alt,
two => \&two,
two_rev => \&two_rev,
});
sub capture {
my $copy = $s;
$copy =~ s/^\s*(.*?)\s*$/$1/s;
}
sub alt {
my $copy = $s;
$copy =~ s/^\s+|\s+$//g;
}
sub two {
my $copy = $s;
s/^\s+//, s/\s+$// for $copy;
}
sub two_rev {
my $copy = $s;
$copy =~ s/^\s+//;
($copy = reverse $copy) =~ s/^\s+//;
$copy = reverse $copy;
}
The output is thus:
Benchmark: running alt, capture, two, two_rev for at least 5 CPU seconds
alt: 5 wallclock secs @ 4163.86/s (n= 21777)
capture: 5 wallclock secs @ 4876.60/s (n= 25846)
two: 6 wallclock secs @ 24841.44/s (n=130666)
two_rev: 5 wallclock secs @ 148958.27/s (n=774583)
Rate alt capture two two_rev
alt 4164/s ---15%-83%-97%
capture 4877/s 17% ---80%-97%
two 24841/s497%409% ---83%
Notice the string this time. It's got embedded whitespace. This causes
the '\s+$' regex to perform HORRIBLY. If you want to see why, try this:
perl -mre=debug -e '"abc def ghi jkl mno" =~ /\s+$/'
Notice that the string "abc def... mno" doesn't end in whitespace, but
that the engine checks for /\s+$/ at each chunk of whitespace. Icky
awful.
--
Jeff "japhy" Pinyan [EMAIL PROTECTED] http://www.pobox.com/~japhy/
RPI Acacia brother #734 http://www.perlmonks.org/ http://www.cpan.org/
what does y/// stand for? why, yansliterate of course.
[ I'm looking for programming work. If you like my work, let me know. ]
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
RE: remove blanks
This is the correct reply.
That's some good stuff.
can you run a benchmark against these, contributed by others on this list,
as well?
$username =~ s/^\s*(.*?)\s*$/$1/;
s/^\s+//, s/\s+$// $username;
thanks,
-rkl
> I ran some benchmarks.
>
> The two-liner outperformed the one-liners by a 10 to 1 ratio. Code and
> results below.
>
>
> Benchmark: timing 10 iterations of OneLine, OneLine2, TwoLines...
> OneLine: 41 wallclock secs (39.30 usr + 0.00 sys = 39.30 CPU) @
> 2544.79/s
> OneLine2: 34 wallclock secs (32.58 usr + 0.00 sys = 32.58 CPU) @
> 3069.56/s
> TwoLines: 3 wallclock secs ( 2.58 usr + 0.00 sys = 2.58 CPU) @
> 38789.76/s
>
>
> use strict;
> use Benchmark;
>
> my $val = "" . ("foo" x 200) . "";
>
> timethese(100_000, {
> 'OneLine' => sub{trimOne($val)},
> 'OneLine2' => sub{trimOne2($val)},
> 'TwoLines' => sub{trimTwo($val)},
> });
>
> sub trimOne {
> my $s = shift;
> $s =~ s/^\s+|\s+$//g;
> die $s unless ($s eq ("foo"x200));
> }
>
> sub trimOne2 {
> my $s = shift;
> $s =~ s/^\s*(.*?)\s*$/$1/g;
> die unless ($s eq "foo"x200);
> }
>
> sub trimTwo {
> my $s = shift;
> $s =~ s/^\s+//;
> $s =~ s/\s+$//;
> die unless ($s eq "foo"x200);
> }
>
>
> -Original Message-
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
> Sent: Monday, September 29, 2003 8:04 PM
> To: [EMAIL PROTECTED]
> Subject: remove blanks
>
>
> Is there a func or a onliner for removing blanks from both ends?
>
> I'm using these:
>
> $username =~ s/^\s+//;
> $username =~ s/\s+$//;
>
> There got to be one out there!
>
> thanks,
> -rkl
>
> --
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
> --
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
RE: remove blanks
That's some good stuff.
can you run a benchmark against these, contributed by others on this list,
as well?
$username =~ s/^\s*(.*?)\s*$/$1/;
$username =~ s/^\s*(.*?)\s*$/$1/;
thanks,
-rkl
> I ran some benchmarks.
>
> The two-liner outperformed the one-liners by a 10 to 1 ratio. Code and
> results below.
>
>
> Benchmark: timing 10 iterations of OneLine, OneLine2, TwoLines...
> OneLine: 41 wallclock secs (39.30 usr + 0.00 sys = 39.30 CPU) @
> 2544.79/s
> OneLine2: 34 wallclock secs (32.58 usr + 0.00 sys = 32.58 CPU) @
> 3069.56/s
> TwoLines: 3 wallclock secs ( 2.58 usr + 0.00 sys = 2.58 CPU) @
> 38789.76/s
>
>
> use strict;
> use Benchmark;
>
> my $val = "" . ("foo" x 200) . "";
>
> timethese(100_000, {
> 'OneLine' => sub{trimOne($val)},
> 'OneLine2' => sub{trimOne2($val)},
> 'TwoLines' => sub{trimTwo($val)},
> });
>
> sub trimOne {
> my $s = shift;
> $s =~ s/^\s+|\s+$//g;
> die $s unless ($s eq ("foo"x200));
> }
>
> sub trimOne2 {
> my $s = shift;
> $s =~ s/^\s*(.*?)\s*$/$1/g;
> die unless ($s eq "foo"x200);
> }
>
> sub trimTwo {
> my $s = shift;
> $s =~ s/^\s+//;
> $s =~ s/\s+$//;
> die unless ($s eq "foo"x200);
> }
>
>
> -Original Message-
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
> Sent: Monday, September 29, 2003 8:04 PM
> To: [EMAIL PROTECTED]
> Subject: remove blanks
>
>
> Is there a func or a onliner for removing blanks from both ends?
>
> I'm using these:
>
> $username =~ s/^\s+//;
> $username =~ s/\s+$//;
>
> There got to be one out there!
>
> thanks,
> -rkl
>
> --
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
> --
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: remove blanks
this looks convenience thanks, -rkl > [EMAIL PROTECTED] wrote: >> Is there a func or a onliner for removing blanks from both ends? >> >> I'm using these: >> >> $username =~ s/^\s+//; >> $username =~ s/\s+$//; >> >> There got to be one out there! > > Doing it in two steps is the way to go. Don't try to make one regex out of > it. > > I usually write it this way, so I can process several variables at once: > >s/^\s+//, s/\s+$// for $foo, $bar, $baz; > > > -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: remove blanks
[EMAIL PROTECTED] wrote: > Is there a func or a onliner for removing blanks from both ends? > > I'm using these: > > $username =~ s/^\s+//; > $username =~ s/\s+$//; > > There got to be one out there! Doing it in two steps is the way to go. Don't try to make one regex out of it. I usually write it this way, so I can process several variables at once: s/^\s+//, s/\s+$// for $foo, $bar, $baz; -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: remove blanks
I ran some benchmarks.
The two-liner outperformed the one-liners by a 10 to 1 ratio. Code and
results below.
Benchmark: timing 10 iterations of OneLine, OneLine2, TwoLines...
OneLine: 41 wallclock secs (39.30 usr + 0.00 sys = 39.30 CPU) @ 2544.79/s
OneLine2: 34 wallclock secs (32.58 usr + 0.00 sys = 32.58 CPU) @
3069.56/s
TwoLines: 3 wallclock secs ( 2.58 usr + 0.00 sys = 2.58 CPU) @
38789.76/s
use strict;
use Benchmark;
my $val = "" . ("foo" x 200) . "";
timethese(100_000, {
'OneLine' => sub{trimOne($val)},
'OneLine2' => sub{trimOne2($val)},
'TwoLines' => sub{trimTwo($val)},
});
sub trimOne {
my $s = shift;
$s =~ s/^\s+|\s+$//g;
die $s unless ($s eq ("foo"x200));
}
sub trimOne2 {
my $s = shift;
$s =~ s/^\s*(.*?)\s*$/$1/g;
die unless ($s eq "foo"x200);
}
sub trimTwo {
my $s = shift;
$s =~ s/^\s+//;
$s =~ s/\s+$//;
die unless ($s eq "foo"x200);
}
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
Sent: Monday, September 29, 2003 8:04 PM
To: [EMAIL PROTECTED]
Subject: remove blanks
Is there a func or a onliner for removing blanks from both ends?
I'm using these:
$username =~ s/^\s+//;
$username =~ s/\s+$//;
There got to be one out there!
thanks,
-rkl
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: remove blanks
On Monday, September 29, 2003, at 07:04 PM, [EMAIL PROTECTED] wrote: Is there a func or a onliner for removing blanks from both ends? I'm using these: $username =~ s/^\s+//; $username =~ s/\s+$//; We could combine those: $username =~ s/^\s*(.*?)\s*$/$1/; Hope that helps. James -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: remove blanks
That is what you want to use. You could do it in a single regex: ## NOT RECOMMENDED - SEE NOTE BELOW ## $username =~ s/^\s+|\s+$//g; The downside is that this is not efficient and actually takes Perl longer to perform the operation. If you want to know why you need to know a little about how Perl stores strings internally. Rob -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Sent: Monday, September 29, 2003 8:04 PM To: [EMAIL PROTECTED] Subject: remove blanks Is there a func or a onliner for removing blanks from both ends? I'm using these: $username =~ s/^\s+//; $username =~ s/\s+$//; There got to be one out there! thanks, -rkl -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
