Am Montag, 2. Mai 2005 03.49 schrieb [EMAIL PROTECTED]:
> John,
>
> the reg exp s <-+> ();
> ^
> was changed to
>
> y/sg//;;
> y/-//d;
> s{\w+} ();
> to exclude the spaces and use y... thx
> why are thre operators y or tr more efficient since these operators do not
> use pattern matching?
Don't know the internal details, but tr needs no backtracking, capturing and
such, it can simply take a char after the other and replace it with another
one. So the implementation of the algorithm must be much much easier.
> Originally I was using an array, actually 2 arrays one with all f string s
> and one with all % strings.
> To me it makes more sense to use a hash b/c each F string needs to be
> pulled with its associative n% string.
> This is why I want to populate a hash of arrays.
Would another strategy be easier and more efficient to fulfill this
requirement?
* If the format of all your data lines is "consistent", you could use split on
\s+ to get the data fields instead of a tr/m cascade.
* Then, if I understand you correctly, you wantto build a hash with % keys and
F... values. This could be done with code like
push @{$lookup_hash{$pct_value}}, $F_value;
eventually, you even want a 2nd level hash with the F field number as key, if
the F values are unique over the whole file and the last field alway begins
with an F.
Or am I overlooking something?
hth, joe
> Below is what is should of read for my population of the hash.
>
> $HoA{$i++} = (split)[-1] if (m/f01(\d+)(\d+%) /gi );
> ciao,
> derek : )
[...]
> [EMAIL PROTECTED] wrote:
> > I was thinking of using a hash of arrays b/c I want to look-up each array
> > by a certain string and that string would the % string.
> > My goal is to populate a hash of some sort with the % string and its
> > associated F string.
> > Here is the data file:
> >
> > 1 2005/01/20 15:39 17 2% -il-o-b----- sg F01000
> > 2 2005/01/20 15:53 14 1% -il-o-b----- sg F01001
> > 3 2005/01/18 09:53 2 0% -il-o-b----- sg F01002
> > 4 2005/02/04 16:41 196 100% -il-o-b----f sg F01003
> > 5 2005/02/05 21:13 305 100% -il-o-b----f sg F01004
> >
> > #!/usr/bin/perl
> > use strict;
> > use warnings;
> > $ENV{"PATH"} = qq(/opt/SUNWsamfs/sbin:/usr/bin:/usr/sbin:/usr/local/log);
> > open (V4, "samcmd v4 2>&1 |" ) or die "unable to open pipe... Broken?
>
> $!";
>
> > my %HoA = ();
> > my $i =0;
> > foreach (<V4>) {
> >
> > s <sg> ();
>
> ^
> The whitespace there will not work for all versions of Perl. Are you sure
> that it works for you?
>
> > s {\-*} ()g;
>
> The hyphen is not special in a regular expression, it does not need to be
> escaped. You are telling the substitution operator to replace all zero
> occurrences of '-' which is unnecessary.
>
> $ perl -Mre=debug -e'$_ = q[ 1 2005/01/20 15:39 17 2% -il-o-b-----
>
> sg F01000]; s{-*} ()g;' 2>&1 | grep -c 'Match successful'
> 55
> $ perl -Mre=debug -e'$_ = q[ 1 2005/01/20 15:39 17 2% -il-o-b-----
>
> sg F01000]; s{-+} ()g;' 2>&1 | grep -c 'Match successful'
> 4
>
> As you can see the regex '-*' matches 55 times while the regex '-+' only
> matches 4 times.
>
> Besides, it would be more efficient to use the transliteration operator.
>
> tr/-//d;
>
> > s {\w+} ();
> >
> > $HoA{$i++} = (split)[-1] if (m/f01(\d+) && (\d+%) /gi );
>
> You are storing the value of $i as the key which starts at 0 and is
> incremented for each line of input so why not just use an array and push
> the
> values onto it? You have included the string ' && ' in your regular
> expression but I don't see that string anywhere in your data? You are
> using
> capturing parentheses in the regular expression but you are not using those
> captured strings anywhere?
>
> > }
> > close (V4) or die "unable to close pipe $!";
> > print "\n";
> >
> > for my $d (keys %HoA) {
> > print "$d: @{ $HoA{$d} }\n";
>
> You are trying to use a scalar value ($HoA{$d}) as an array which strict
> should complain about.
>
> > }
>
> John
[...]
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
<http://learn.perl.org/> <http://learn.perl.org/first-response>
