Regex with xml tag

**Vodka** · 04-02-2010

Hello,
I work on a String containing XML tags. I try to recognize certain patterns:

Code:

t <ind> test </ ind> with a <ind> </ ind> More

Here my goal is to extract:

Code:

t with result

To isolate the text unmarked.
I use the following regexp: <ind> .* </ ind>
AC works well, but I just noticed that it works too well because for the example above if I return

Code:

Following t

Because of

Code:

t<ind>test </ ind> with a <ind></ ind> Following

So, how can I use the Regexp effectively with xml tags, so that I can extract data from it.

**Windowed** · 04-02-2010

Hello,
Note that the regex is not often the right tool to parse the xml and they work well only for the most simple xml. In your case, try:

Code:

<ind> .*?</ ind>

I think this is the suitable option for you. And as i said earlier regex is not a good ideas to parse a xml. So, it is better if you find a other way round. Finally it is up to you, you can use what ever you want.

**Vodka** · 04-02-2010

Hi,

Code:

<ind> .*?</ ind>

Yes it works better but still not it for this code, it is

Code:

t<ind> test </ ind> with <ind> a <ind> two </ ind></ ind> More

If regexp is not the best way to operate, what is it? using languages such XPATH? I have no idea about the others, if you know one which best fits here, then you can post or give me the link to the article or documentation on it.

**cyber-noob** · 04-02-2010

Hello,
I think you should limit, i.e make sure that XML remains compatible with a regex analysis, there is no need for a kind of historical reasons (the classic separation lex / yacc or syntax / grammar , there is no grammar). Here the overlap is very difficult to treat in regex. If I want to do anyway, so I do a regex with the OR group match beginning and end of the ground, I walk my chain with a find and I react as I find a beginning or an end; to I end the global as you want, but it only works if no comments or other in XML.

**Vodka** · 04-02-2010

Hi,
Yes that's how I proceed to a number of tags, I get a ground opening and closing. I look down the list returned by both find () to find the extremism. Then by recursion, I take the string contained between the two extrema and I again. I'll see to apply this principle here at first (because I'm in a hurry to make this work ), But I will look at ways to use an XML API to locate the positions of tags in a String.

**SoftWore** · 04-02-2010

Hi,
Here is a regex that would please you (perl-compatible, but it should go in java)

Code:

<(\ w +)>(.*?)</\1>

It does not pass on the stroke of nested tags, but the rest is ok. You can not do recursion, to my knowledge, with regex, so what you're looking to do is impossible. But in fact, an XML parser is probably easier.