Difficulties with split

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

Replies are listed 'Best First'.
Re: Difficulties with split by almut (Canon) on Apr 27, 2009 at 23:06 UTC
If I'm understanding you correctly, you want `@vals = unpack "(A2)*", $in;` [download]	[reply] [d/l]
Re^2: Difficulties with split by ikegami (Patriarch) on Apr 27, 2009 at 23:48 UTC
That doesn't work with strings internally encoded as UTF8 before 5.10.0 (and not at all before 5.8.0). If you're working with arbitrary characters, if you don't to mess with `utf8::downgrade`, or if you are worried about backwards compatibility, you can use the following instead: `@vals = $in =~ /(..)/sg;` [download]	[reply] [d/l] [select]
Re^2: Difficulties with split by transiency (Sexton) on Apr 27, 2009 at 23:10 UTC
W00t, that is indeed what i wanted, thank you almut. :) foreach(@the_wicked){sleep(0);}	[reply]
Re: Difficulties with split by runrig (Abbot) on Apr 27, 2009 at 23:08 UTC
Think about it. You are splitting ON any two characters. If you were splitting on, say, a comma (","), would you expect commas to end up in the result? Some options, ~~one, use a zero-width look behind assertion~~: `split /(?<=..)/, $in; # johngg is right...bad answer` [download] or capture what you're splitting on, and filter out empty elements: `grep { length($_) > 0 } split /(..)/, $in;` [download] or do like almut says above and use unpack :-)	[reply] [d/l] [select]
Re^2: Difficulties with split by johngg (Canon) on Apr 28, 2009 at 13:51 UTC
`split /(?<=..)/, $in;` I don't think that is going to do quite what you expected. It will `split` the string at every point that is preceded by two characters. `$ perl -le ' -> $str = q{abcdefg}; -> @bits = split m{(?<=..)}, $str; -> print for @bits;' ab c d e f g $` [download] It seems that `split` is not perhaps the handiest tool for this task. Cheers, JohnGG	[reply] [d/l] [select]
Re^2: Difficulties with split by transiency (Sexton) on Apr 27, 2009 at 23:22 UTC
That is a good point, i forgot split removes the delimitter heh, and should've tried capping it. I've never heard of a look behind assertion, it sounds interesting... Thanks runrig :) foreach(@the_wicked){sleep(0);}	[reply]
Re: Difficulties with split by jwkrahn (Abbot) on Apr 28, 2009 at 00:43 UTC
Or you could use a regular expression to do what you want: `my @vals = $ARGV[ 0 ] =~ /../g; print @vals;` [download]	[reply] [d/l]
Re^2: Difficulties with split by przemo (Scribe) on Apr 28, 2009 at 14:09 UTC
Don't forget about the last char in strings of odd length: `my @vals = $ARGV[ 0 ] =~ /..\|.$/g; print @vals;` [download]	[reply] [d/l]
Re^3: Difficulties with split by jwkrahn (Abbot) on Apr 28, 2009 at 14:55 UTC
OK: `my @vals = $ARGV[ 0 ] =~ /..?/g; print @vals;` [download]	[reply] [d/l]
Re: Difficulties with split by perliff (Monk) on Apr 28, 2009 at 09:04 UTC
my 2 cents... `#!/usr/bin/perl use warnings; $in = $ARGV[0]; @vals = split(/\.\./, $in); print @vals;` [download] perliff ---------------------- "with perl on my side"	[reply] [d/l]
Re^2: Difficulties with split by johngg (Canon) on Apr 28, 2009 at 22:12 UTC
A full stop is a regular expression metacharacter meaning match any character (with caveats, see perlretut, perlre). However, a backslash escapes the special meaning of a metacharacter so that `\.` matches a literal full stop. That means you are spliting on two literal full stops, which is not quite what was wanted. `$ perl -le ' > $in = q{abcdefg}; > @vals = split m{\.\.}, $in; > print for @vals;' abcdefg $ perl -le ' > $in = q{abc..defg}; > @vals = split m{\.\.}, $in; > print for @vals;' abc defg $` [download] I hope this is of interest. Cheers, JohnGG	[reply] [d/l] [select]