Regex pattern for

Question

I am trying to apply regex pattern like this.

I want to apply pattern like this.

<a attributes="some set of attributes"><img attributes="some set of attribtes"/></a>

Rules:

    <a> tag with attributes followed by <img> with attributes.

Sample Valid Data:

        <a xlink:href="some link" title="Image" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns="http://www.w3.org/1999/xhtml">
            <img  alt="No Image" title="No Image" xlink:href="soem path for image" xlink:title="Image" xmlns="http://www.w3.org/1999/xhtml" xmlns:xlink="http://www.w3.org/1999/xlink" />
        </a>

Invalid:

    <a>data<img/></a>--Data Present, no attributes
    <a><img>abcd</img></a>--data Present, No attributes
    <a><img/></a>---No attributes

Can any one suggest how to write pattern for this.

Thank you.

Why do you insist on a regex? [Use a real parser](http://stackoverflow.com/a/1732454/1002469) — Francis Avila, Jun 29 '12 at 05:07
You are about the 1701th person asking something similar. Search, read, post. In that order. — PointedEars, Jun 29 '12 at 05:08
@PointedEars pointed ears... 1701... coincidence... I think not :D — Bohemian, Jun 29 '12 at 05:19
@Bohemian Correct :) But sorry to all native speakers for "1th". — PointedEars, Jun 29 '12 at 07:30
Trying to parse HTML with regex [induces madness](http://stackoverflow.com/questions/1732348/regex-match-open-tags-except-xhtml-self-contained-tags?page=1&tab=votes#tab-top). — Avaq, Jun 29 '12 at 10:56
@Avaq Contrary to popular belief, it is possible, but only with extreme care. — PointedEars, Jun 29 '12 at 21:27

Francis Avila · Answer 1 · 2012-06-29T05:41:06.947

You can do this in a completely bulletproof manner with XPath:

//*[local-name()='a' and count(@*)>0 and *[local-name()='img' and count(@*)>0] and count(.//*)=1 and normalize-space(.)='']

This selects all elements with a local name of 'a' which have no non-significant text content, attributes, and a single 'img' element with attributes.

However, since your example code is clearly XML with namespaces and all, perhaps you can reformulate your question to say what your overall task is instead of "what regex should I use". At the very least it seems that perhaps you should be paying attention to those namespaces instead of treating namespace declarations as attributes.

For example, maybe what you really mean is this?

//xhtml:a[@xlink:href and xhtml:img[@xlink:href] and count(.//*)=1 and normalize-space(.)='']

Regex pattern for

1 Answers1