Beefy Boxes and Bandwidth Generously Provided by pair Networks
Pathologically Eclectic Rubbish Lister

(Ovid - don't use regexes for parsing) Re(2): Parsing XML

by Ovid (Cardinal)
on Jun 21, 2001 at 22:19 UTC ( #90479=note: print w/replies, xml ) Need Help??

in reply to Re: Parsing XML into a simple hash
in thread Pasring XML into a simple hash

Your solution does not work. Initially, it will appear to work against his data set, but XML start and end tags don't have to appear on the same line. If that happens, your regex will break because the dot metacharacter doens't match the newline. Adding the /s modifier allows the dot to match, but then, because your match is greedy, it still breaks:

#!/usr/bin/perl use strict; my @buf = <DATA>; for my $i ( 0 .. $#buf ) { if ($buf[$i] =~ s/^\s*<jobnumber>(.*)<\/jobnumber>\s*$/$1/s) { $buf[$i+1] =~ s/^\s*<location>\s*(.*)\s<location>\s*$/$1/s; # if your tags are really like this &process($buf[$i],$buf[$i+1]); } } sub process { my ($jobnumber,$location) = @_; print "Found a job $jobnumber in $location.\n"; # do something } __DATA__ <posts> <post> <jobnumber> 1234 </jobnumber> <location>Somecity, NJ</location> </post> <post> <jobnumber>87922</jobnumber> <location>Othercity, AK</location> </post> </posts>

See Death to Dot Star! for the explanation of why your regex fails (and for some excellent examples of how I have screwed up regexes on delimited text).

Use a parser for data like this. Regexes, while I love them, are for matching data, not parsing it.

As for your 'related note', it doesn't work because you have (.?) in your code. The dot/question mark makes you match one character and have that match optional. It's equivalent to (.{0,1}).


Join the Perlmonks Setiathome Group or just click on the the link and check out our stats.

Replies are listed 'Best First'.
Re: (Ovid - don't use regexes for parsing) Re(2): Parsing XML
by mattr (Curate) on Jun 22, 2001 at 14:05 UTC
    Thanks Ovid, you're right and I'll reread that article!

Log In?

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://90479]
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others pondering the Monastery: (5)
As of 2020-05-26 18:22 GMT
Find Nodes?
    Voting Booth?
    If programming languages were movie genres, Perl would be:

    Results (150 votes). Check out past polls.