ugly code

fubber has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: ugly code by Limbic~Region (Chancellor) on Jul 07, 2007 at 22:29 UTC
fubber, Ultimately you are going to have to loop even if you don't want to. That said, it seems that you have a list of characters that are being translated to other characters. Perhaps you want tr instead. `tr/abc/efg/ for @subtitles;` [download] Cheers - L~R	[reply] [d/l]
Re: ugly code by almut (Canon) on Jul 07, 2007 at 22:29 UTC
If it's single chars being replaced by single chars, you might also use `tr///` `map tr/şţŞ/stS/, @subtitle;` [download]	[reply] [d/l] [select]
Re: ugly code by Joost (Canon) on Jul 07, 2007 at 22:24 UTC
Since `$_ =~ s///` is exactly equivalent to `s///` you can do `foreach ( @subtitle ) { s/Âş/s/; s/Ăž/t/; s/ÂŞ/S/; }` [download] I don't know what you mean in the rest of your post, though, I assume you mean something like: `s/Âş/s/, s/Ăž/t/, s/ÂŞ/S/ for @subtitle;` [download] note: all my code is untested. "What should it profit a man, if he should win a flame war, yet lose his cool?"	[reply] [d/l] [select]
Re: ugly code by jZed (Prior) on Jul 07, 2007 at 22:23 UTC
`@subtitles = map { s/ş/s/; s/ţ/t/; s/Ş/S/; $_; } @subtitles;` [download]	[reply] [d/l]
Re: ugly code by parv (Parson) on Jul 08, 2007 at 03:24 UTC
Yet another way to do it is to use a hash (which in OP's case seem out of proportion) ... `my $string = 'bulky axe'; my %map = ( 'a' => 2 , 'b' => 4 ); $string =~ s/$_/$map{ $_ }/g for keys %map;` [download] And, Python may have a solution for your "white, ugly braces" problem.	[reply] [d/l]
Re^2: ugly code by bart (Canon) on Jul 08, 2007 at 16:10 UTC
Better: `$string =~ s/(a\|b)/$map{ $1 }/g;` [download] You may be able to construct the pattern out of the hash keys: `$pattern = join '\|', map quotemeta, keys %map; $string =~ s/($pattern)/$map{ $1 }/go; # /o only if the pattern never + changes` [download] For not too recent perls, you get a better performance if you build the pattern out of the keys with a module like Regexp::Assemble or Regex::PreSuf; `use Regex::PreSuf; my $pattern= presuf(keys %map); $string =~ s/($pattern)/$map{ $1 }/go;` [download] Allegedly, the most modern perl will optimize it for you, and possibly even with better (faster) results. update (July 9) ysth sent me a msg with a caution for the case where one of the hash keys is a prefix of another key, for example: ('foo', 'food'). In that case, to make sure the longest key is matched, you should sort the keys when constructing the pattern so the longest key comes first. ysth proposed: `$pattern = join '\|', map quotemeta, sort { length($b) <=> length($a) } + keys %map;` [download] which will work, but so will this: `$pattern = join '\|', map quotemeta, sort { $b cmp $a } keys %map;` [download] because with two strings where one is a prefix of another, the longer one will compare as "greater" than the shorter one. You need not worry when using Regex::PreSuf because it'll always attempt to match the longest string, first.	[reply] [d/l] [select]
Re^3: ugly code by johngg (Canon) on Jul 08, 2007 at 22:15 UTC
I don't know whether this is better or not but I tend to use the quotishness of compiled regexen with a localized list separator rather than `join`. `my $rxPattern; { local $" = q{\|}; $rxPattern = qr{(?x) (@{ [ map { quotemeta } keys %map ] }) }; } $string =~ s{$rxPattern}{$map{$1}}g;` [download] I'd be interested to know if this is a sensible approach. Cheers, JohnGG	[reply] [d/l] [select]