search & replace regular expression help

[ni_c]

Hi, im looking for a regular expression to replace html rules without changing the content inside

so if a documents sets italic rules like

<span class="italic">example one</span>
<span class="italic">example two</span>
<span class="italic">example three</span>

am looking to change them to

<i>example one</i>
<i>example two</i>
<i>example three</i>

and only know a regex search parameter of

<span class="italic">(.*?)</span>

will find all instances, but dknow the replace equivalent to keep the content

so assuming (.*?) is a search only parameter, whats the replace equivalent?

[jbacelar]

<i>\1</i>

And, I recommend: http://www.regular-expressions.info/tutorial.html

[eschwartz]

The parentheses form a capture group. You can reuse a capture group in a replacement, with an escaped number -- e.g. "\1" is the replacement for the first capture group.


This is more of a general regex question. You may want to bookmark http://regular-expressions.info (my favorite regex tutorial and reference) for quick reference instead of opening a thread here

[mikapanja]

Use <i>\1</i> for replace.
If covering multiple instances, go on with \2, \3 etc.
Don't forget to turn the REGEX mode on.

[ni_c]

thanks guys, had been confused by \1, last time i tried it, thought it just replaced everything with \1 for some reason. Works now, cheers, and thanks for the links, will read up on this

[eschwartz]

Hah, we all piled on at once

[davidfor]

Alternatively, you can use Diap's Editing Toolbag plugin. One of its functions is will do exactly this type of replacement.

[Tex2002ans]

Originally Posted by davidfor

Alternatively, you can use Diap's Editing Toolbag plugin. One of its functions is will do exactly this type of replacement.

I second this. Depending on how clean the input is, this would be more robust than a Regex solution.

Here is an image showing you the specific settings you would use when using Diap's Editing Toolbag to go from "<i>Example</i>" to "<span class="italic">Example</span>":

[theducks]

Originally Posted by ni_c

thanks guys, had been confused by \1, last time i tried it, thought it just replaced everything with \1 for some reason. Works now, cheers, and thanks for the links, will read up on this

If it DID that,
you forgot to enclose the keep this part in a positional # matching ( )
You see the \3, then the were only 2 ( ) in the search where you replace was
\1 foo \2 bar \3

[ni_c]

good stuff, will install that, see how it goes,

fwiw, the regex expression worked beautifully for what i actually wanted, converting destination footnote links into dual direction ones

first gave footnotes ids like id="f-1", id="f-20" etc

then converted destination placeholders like
<a id="filepos789849"></a>20.

using regex search expression
"></a>(.*?)

and replace expression
" href="Chapter-01.html#f-\1">\1</a>

resulting in
<a id="filepos789849" href="Chapter-01.html#f-20.">20.</a>

only needed to hit "replace and find" till the end of each chapter, change chapter numbers and continue, very quick, then just used normal mode to find ."> and replace with "> (to correct the ids)

[theducks]

Originally Posted by ni_c

good stuff, will install that, see how it goes,

fwiw, the regex expression worked beautifully for what i actually wanted, converting destination footnote links into dual direction ones

first gave footnotes ids like id="f-1", id="f-20" etc

then converted destination placeholders like
<a id="filepos789849"></a>20.

using regex search expression
"></a>(.*?)

and replace expression
" href="Chapter-01.html#f-\1">\1</a>

resulting in
<a id="filepos789849" href="Chapter-01.html#f-20.">20.</a>

only needed to hit "replace and find" till the end of each chapter, change chapter numbers and continue, very quick, then just used normal mode to find ."> and replace with "> (to correct the ids)

I would pick a safer capture pattern to avoid strays

(\d+)

Only a Integer Number (any amount digits)

[Thom*]

This is so simple you don't need a plugin to accomplish the task.
The only thing I would change, in order to avoid getting messed up with nested <span> tags, is to change the search string.

Search: <span class="italic">(((?!<span).)*?)</span>
Replace: <i>\1</i>

Consider nested span tags:
<span class="italic">Merry <span class="bold">Christmas </span>Everyone.</span>

Search: <span class="italic">(.*?)</span>
Replace: <i>\1</i>

Would result in:
<i>Merry <span class="bold">Christmas </i>Everyone.</span>

The search/replace that I described does a look-ahead and would not select the span because it has a nested span. You would have to address the "span class="bold" first, then rerun the "span class="italic" to modify it.

[eschwartz]

Worth noting -- much as I am a fan and advocate of lookarounds, the plugin saves you from having to do that.


For many people it's worth it for the convenience.

[ni_c]

Originally Posted by theducks

I would pick a safer capture pattern to avoid strays

(\d+)

Only a Integer Number (any amount digits)

thanks theducks, that one works perfectly

[phossler]

Originally Posted by ni_c

thanks guys, had been confused by \1, last time i tried it, thought it just replaced everything with \1 for some reason. Works now, cheers, and thanks for the links, will read up on this

Sometimes I get confused and use the tag-ending / instead of the group replacement \

That DOES make a mess