Hello,

I have a very large XML file, containing some XHTML. In this file, there are hundreds of <p>s which I want to change to <table>s. All of the <p>s I want changed have a specific ID, so it seems like it should be easy to replace them, but I can only get 36 of them. I am using HomeSite 5.2.

Find what:

<p id="21">Note:([A-Za-z0-9_ ><.="']*)</p>

Replace with:

<table rows="2" cols="2" width="8744"><tr rowheight="0"><td width="630" bordercolors="65793"><p id="4"><fref id="13453"/></p></td><td width="8114" bordercolors="65793"><p id="4">\1</p></td></td></table>

If I search for <p id="21">Note:, I get 186 hits. If I use the above replacement, it works great, but for only 36 hits. Some (but not all) of the items not found have additional coding in them, but I would think that the >< in the Find what string would take care of them. Unfortunately, I barely have a clue what I'm doing, and I've taken the [A-Za-z0-9_ ><.="']* from somebody else's example without understanding it. I'm not sure I've even titled this request correctly. Thanks so much for any suggestions.

We can't see your file, please post it to a site such as pastebin.com if you want us to test with it.

  [A-Za-z0-9_ ><.="']*     any character of: 'A' to 'Z', 'a' to 'z',
                           '0' to '9', '_', ' ', '>', '<', '.', '=',
                           '"', ''' (0 or more times (matching the
                           most amount possible))

---

Thanks for the quick reply. To avoid sharing this huge amount of data, I've pasted a representative sample into pastebin here: http://pastebin.com/m7ad9ecee

If you are working with XML then XSLT might be a better option than a regex.

also your table markup has a typo, no closing tr but extra closing td.

---

Michael

"In theory, theory and practice are the same. In practice, they are not."
Albert Einstein

mash made a good point on not using regex if you can use another method designed specifically for XML, but this regex match pattern worked for me based on your source text:

<p id="21">.*?</p>

http://livedocs.adobe.com/coldfusion/5.0/Using_HomeSite/language5.htm

---

Thanks for the tip. I guess I will have to learn XSLT. I was hoping there'd be an easier option, since I felt like I was close with regex.

The match pattern you suggested seems so elegant, just what I was hoping for, but in HomeSite it generated an error 17, and I can't find any searchable reference online that describes errors by number. Then I realized that it wouldn't do what I need anyway, because I need to match a pattern at the beginning of some random content, and the pattern at the end (</p>). All the while replacing the content when I change the pattern, so I need a group in the search pattern. I tried this:

<p id="21">Note:(.*?)</p>

and replaced with the same as before (after fixing the </td></td> error, thanks mash). But it found the first instance of  <p id="21">Note:, and added the entire document to the group, and then replaced the last </p> in the document, skipping over other instances of <p id="21">Note: content</p>.

Your next attempt might be:

<p id="21">Note:(?:(?!</p>).)*</p>

---

I appreciate your continued suggestions. That last one generated no hits, whether doing the Find and Replace or a Find. Faced with the (for me) daunting task of learning XSLT, I messed with it a bit to see what would happen.

If I try this: <p id="21">Note:(?:(?!</p>).)*

I get four hits for "<p id="21">Note:" when Finding, but doing my find and replace generated Regular expression error 0:

I guess I'm just flailing at this point.

Application issues are somehow interfering with your operation, see the pattern in use with your supplied text here:

http://www.myregextester.com/?r=306

---

What application would you recommend for doing a find and replace of this nature on a very large xml file (or set of xml files)?

Not necessarily XML-specific apps, but here are two options I have seen:

http://tools.tortoisesvn.net/grepWin

http://www.powergrep.com/

PowerGREP have a larger feature set but it is a commercial application.

I don't have experience with XML-specific find/replace apps so that's the best I can recommend at this point.

---

grepWin is really easy to use. Thanks for the link. When I tried this:

 Find: <p id="21">Note:(?:(?!</p>).)*</p>

 Replace with: <table rows="2" cols="2" width="8744"><tr rowheight="0"><td width="630" bordercolors="65793"><p id="4"><fref id="13453"/></p></td><td width="8114" bordercolors="65793"><p id="4">\1</p></td></tr></table>

it found every instance, but killed the content. The \1 was gone, and there was nothing there. But if I do this:

 Find: <p id="21">Note:([A-Za-z0-9_ ><.="']*)</p>

 Replace with: <table rows="2" cols="2" width="8744"><tr rowheight="0"><td width="630" bordercolors="65793"><p id="4"><fref id="13453"/></p></td><td width="8114" bordercolors="65793"><p id="4">\1</p></td></tr></table>

That works fine, but only for 36 of the 187 instances. I feel like I'm closer, but the regex is not quite right. Meanwhile, I've been reading up on doing find-replace with XSLT, and XSLT uses regex for this. So it appears that regex is my best hope after all.

Thanks all for bearing with me.

Find: <p id="21">Note:((?:(?!</p>).)*)</p>

Your first pattern was missing capture group 1.

---

Cool, thank you so much, this is now working in grepWin.