Adding text if missing (Perl REGEXP)

Question

Using a Perl replace statement (s//) I'd like to modify some XML by adding an attribute to some items if missing. Here's an example :

...
<car color="red" owner="john">...</car>
<car color="green" age="3">...</car>
...

Let's say I'd like to add a default owner to ownlerless cars. I tried the following without success :

s/(<car[^>]*)(?!owner="[^"]*")(.*>)|$1 owner="steve"$2/iUg

Any help appreciated.

possible duplicate of [In Perl, how can I change an element in an XML file without changing the format of the XML file?](http://stackoverflow.com/questions/1327720/in-perl-how-can-i-change-an-element-in-an-xml-file-without-changing-the-format) — Schwern, Jan 22 '12 at 17:53
avoid using regexes to parse / change XML or HTML documents. that's a general recommendation. — snoofkin, Jan 22 '12 at 23:00

score 5 · Accepted Answer · answered Jan 22 '12 at 17:11

5

While you may be able to get away with a regex, an XML parser is always recommended.

Perl/CPAN offerings include:

XML::Simple

Treats your XML document like a Perl data structure
XML::Parser

An old-timer
XML::SAX

For SAX parsing needs
XML::Twig

For memory-efficient processing of XML documents in chunks
XML::LibXML

My personal favorite

answered Jan 22 '12 at 17:11

Zaid

36,680
16
86
155

1

You missed out on an opportunity to link to [the most awesome SO answer ever](http://stackoverflow.com/a/1732454/301832). – Donal Fellows Jan 22 '12 at 21:15

score 1 · Answer 2 · answered Jan 22 '12 at 19:59

1

I usually use XML::XSH2 for XML manipulation. Your problem can be solved easily in it:

for //car[not(@owner)] set @owner "steve" ;

answered Jan 22 '12 at 19:59

choroba

231,213
25
204
289

user1146332 · Answer 3 · 2012-01-22T20:57:12.147

0

you can try this one

#!/usr/bin/perl -w

$str = <<EOF;
<car color="red" owner="john">.*#?§<car color="pink"> </car>.*#?§</car>
<car color="green" age="3">.*#?§</car>
<car owner="bill"></car>
EOF

$str =~ s/(<car(?![^>]+?owner))([^>]*?)>/$1$2 owner="steve">/g;

print $str;

best regards

PS: most likely it's a typo but anyway, the vertical bar in your regex must be replaced by a slash

edited Jan 22 '12 at 20:57

answered Jan 22 '12 at 17:46

user1146332

2,630
1
15
19

1

There's no need for the empty while loop. `/g` already makes sure it runs over all elements of `$str`. The loop just makes it run twice, and the second time around it doesn't match anything so it returns 0 (the number of matches). This could easily become an infinite loop, for example: `s{foo}{foofoo}g`. – Schwern Jan 22 '12 at 17:56
@Schwern thanks for your comment. I remember:`m//g != s///g` best regards – user1146332 Jan 22 '12 at 20:56

Adding text if missing (Perl REGEXP)

3 Answers3