http://qs321.pair.com?node_id=814374


in reply to Re: Perl/Ruby comparison
in thread Perl/Ruby comparison

His Perl could have been tightened up considerably.

Boy, is that a call to action! I can't believe I'm the first in 4 years to have taken it up …. Here's a more-or-less faithful copy:

use strict; use warnings 'all'; use constant { EMAIL => 17, CONTACTME => 27, SKUTITLE => 34 }; my @records; # Arrays are initialised empty; no need to do it explicit +ly. push @records, [ split "\t" ] for <>; do { tr/"//d for @$_ } for @records; my @contact_records; do { push @contact_records, [ @$_[ SKUTITLE, CONTACTME, EMAIL ] ] if $ +_->[CONTACTME] == 1 } for @records; { local $, = "\t"; local $\ = "\n"; print @$_ for sort { $a->[0] cmp $b->[0] } @contact_records; }

To me, this task suffers, rather than benefits, from the names, and is much better suited for a Unix-style pipeline; fortunately, Perl is good at those (although I resort to substr to avoid the to-me unsatisfying s///; $_):

use strict; use warnings 'all'; use constant { EMAIL => 17, CONTACTME => 27, SKUTITLE => 34 }; { local $, = "\t"; local $\ = "\n"; print @$_ for sort { $a->[0] cmp $b->[0] } grep { $_->[1] } map { chomp; [ map { substr $_, 1, -1 } ( split "\t" )[ SKUTITLE, CONTACTME, EMAIL ] ] } <>; }
I've replaced the test for whether the CONTACTME field == 1 (not eq "1", ugh) with a test for whether it's true, since that's almost certainly what was meant; but it's easy to change it back.

Finally, note that, since the CONTACTME field is not carrying any useful information, we might as well avoid printing it; so we could define

sub snip_or_toss { $_[CONTACTME] ? @_[ SKUTITLE, EMAIL ] : () }
replace the outer map by map { chomp; [ map {...} snip_or_toss split "\t" ] }, and then get rid of the grep.