Re: [xsl] match string

[Anton Triest <anton@xxxxxxxx>]

Zsolt Szabs wrote:

> I have modified my template, because like you said I had lost the 
> markup in para[1]
>
> 1.
> <xsl:template match="chapter/section[1]/para[1]/*[1]/text()[1]" 
> priority="1">
>
> 2.
> <xsl:template match="para">
>
> but now the output of the first paragraph looks like>
> <p> <p> <b> Three words are </b> here </p> </p>
>
> it is because I have a template for the first paragraphs first tag text()
> and a template for every paragraph, with matches para[1] too

Yes (like David says): in the template matching text() you don't need a 
p element node anymore.

> how can I solve this problem?
>
> On Oct 19, 2004, at 3:04 PM, David Carlisle wrote:
>
> > Also, in this case you probably want to match on the first text node of
> > the first para rather than the para itself as in
> >
> > <xsl:template match="chapter/section[1]/para[1]//text()[1]" 
> > priority="1">
> >
> > As you have it, you are doing string operations on the entire paragraph
> > text which means that any element markup within the paragraph will be
> > lost.
There's another problem. I thought "//text()[1]" would give "the first 
text node (inside that first paragraph) however deeply nested in any 
markup". But instead, it seems to return the first text node of *all* 
the nested markup. Example:

   <input>
       <para>A paragraph without any markup</para>
       <para>A paragraph with some <i>markup</i> inside</para>
       <para>A paragraph with some <b><i>nested</i> markup</b></para>
       <para><em>This is a special case</em>: paragraph starts with 
markup</para>
       <para><em>This</em> is difficult: only the first word has 
markup</para>
   </input>

   <output>
       <para><first>A paragraph without </first>any markup</para>
       <para><first>A paragraph with 
</first>some<i><first>markup</first></i> inside</para>
       <para><first>A paragraph with 
</first>some<b><i><first>nested</first></i><first>markup</first></b></para>
       <para><em><first>This is a </first>special case</em><first>: 
paragraph starts </first>with markup</para>
       <para><em><first>This</first></em><first>is difficult: only 
</first>the first word has markup</para>
   </output>

Zsolt, I see that you changed this to "/*[1]/text()[1]" but that gives

   <output>
       <para>A paragraph without any markup</para>
       <para>A paragraph with some <i><first>markup</first></i> 
inside</para>
       <para>A paragraph with some 
<b><i>nested</i><first>markup</first></b></para>
       <para><em><first>This is a </first>special case</em>: paragraph 
starts with markup</para>
       <para><em><first>This</first></em> is difficult: only the first 
word has markup</para>
   </output>

> your solution works, but it is to complicated for me. I need time to 
> exactly understand it.

Don't worry Zsolt, it's getting complicated for me too! :-) It think I 
also need some time to work this out

Best regards
Anton