in reply to Ordering hash replacements to avoid clobbering things (update chaining)
•Update: I've retracted this post, it doesn't apply to the situation (I missed the part about it being db records that need to be updated)
This stuff below is only useful if you need to do the substitutions in a string.
If the search-string needs to occur as a word by itself, use \b, like:
Final note: this code assumes the keys don't contain any odd chars like regex metachars. Since your example shows all-uppercase abbreviations, so apparently this isn't a problem. If it is, then you'll need to look into using quotemeta.
Produces the output:use strict; use warnings; my %replace = qw( COM SEC MOC COM SEC COM CO MOC ); my $searchpat = join '|', sort { length $b <=> length $a } keys %repla +ce; $searchpat = qr/$searchpat/; # optional, dunno if it speeds up things my $example = "FOO CO COM BAR MOC SEC"; $example =~ s/($searchpat)/$replace{$1}/g; print "$example\n";
As you can see, it also handles circularity without problem; COM <-> SEC.FOO MOC SEC BAR COM COM
Note that I'm sorting the keys on length to make sure longer patterns match first (otherwise "COM" would become "SECM").
If the search-string needs to occur as a word by itself, use \b, like:
which prints:my $example = "MOC.COMPUTER.COM.FOO / SEC.SECT / CO"; $example =~ s/\b($searchpat)\b/$replace{$1}/g; print "$example\n";
COM.COMPUTER.SEC.FOO / COM.SECT / MOC
Note that in this case sorting on the length of the key is not necessary, so you can simply do my $searchpat = join '|', keys %replace;
Final note: this code assumes the keys don't contain any odd chars like regex metachars. Since your example shows all-uppercase abbreviations, so apparently this isn't a problem. If it is, then you'll need to look into using quotemeta.
|
---|
Replies are listed 'Best First'. | |
---|---|
Re:x2 Ordering hash replacements to avoid clobbering things
by grinder (Bishop) on Feb 26, 2003 at 16:17 UTC | |
by xmath (Hermit) on Feb 26, 2003 at 16:33 UTC |
In Section
Cool Uses for Perl