Match Everything Between Tags That Contain Key

Question

I'm trying to find a (sed style) regex that will match every instance of the word "HAWK" and the closest surrounding item tags, i.e. <item> ... HAWK ... </item>, where the ellipsis may be text or other tags (but not the item tag).

So far I've tried lazy match-all: <item>(.*?)HAWK(.*?)<\/item>, and find that this works well for catching everything between HAWK and the closing item tag, but matches over many nested opening <item> tags, and so winds up getting too much.

I think using look-behind might help but I've had problems getting this to work properly also. Any help would be much appreciated.

There are lots of xml parsers, please don't rewrite them poorly from scratch. — Aaron Hall, Feb 05 '14 at 03:21
s/(.*?)HAWK(.*?)//g. But as I explained this does not achieve what I want. Why does this matter? It's the regex that I'm concerned with and I don't think the logic is going to change because I'm using a UNIX flavour. — Ryan, Feb 05 '14 at 03:27
[Obligatory link](http://stackoverflow.com/q/4231382/471272). — tchrist, Jun 08 '14 at 20:01

score 0 · Accepted Answer · answered Feb 05 '14 at 05:06

0

In order to find the "closest tag" using sed style expressions, you can try

/<item>[^>]*HAWK[^<]*<\/item>\)/

This works in 'regular sed'. Basically, by matching anything that is not a closing bracket before HAWK, and anything that is not an opening bracket after HAWK, it 'simulates' the non-greedy quantifier .*?.

But it is usually not a good idea to try to parse XML etc with regex - a 'real parser' is much more robust. But hey, you asked.

answered Feb 05 '14 at 05:06

Floris

43,828
5
63
112

1

Yep, I did, and I really appreciate you answering my original question anyway. – Ryan Feb 05 '14 at 05:20

Match Everything Between Tags That Contain Key

1 Answers1