Re: [xsl] testing element's contents

Subject: Re: [xsl] testing element's contents
From: Joerg Heinicke <joerg.heinicke@xxxxxx>
Date: Thu, 03 Oct 2002 19:07:40 +0200
Hi Nathan,

you seem again to be wrong here. Tha value of a node is the concatenated string of every descendant text node. So

<xsl:template match="p[not(normalize-space(.))]">

will not match on <p><strong>this is bold text</strong></p>.



Nathan Shaw wrote:
Thanks Wendell. Thinking it over, your simple template

<xsl:template match="p[not(normalize-space(.))]">

will not work, as much as I wish it would, as I am
allowing other elements such as <strong> and <em> and
I could get passed a node set like this

<p><strong>this is bold text</strong></p>

OR this

<p><em>this is bold text</em></p>

For both of those cases, they would need to get
wrapped in a p tag in the output.

So, I do need to specifically look to see if the p
element's child nodes are made up of either

1) only a Media element
2) only a Media element wrapped by a Link element

If either of those cases is true, I do not want to
wrap the output in a p tag. Otherwise, I do.

I hope that makes it a bit clearer.


--- Wendell Piez <wapiez@xxxxxxxxxxxxxxxx> wrote:

Hi Nate,

At 11:08 AM 10/3/2002, you wrote:

hi all,

Thanks to everyone who gave me advice on my


post about images and captions. Now, I have another
tricky lil' xslt problem. If a paragraph element
contains ONLY a media element OR a media element
surrounded by a link element and nothing more


no other nodes, be they text or not), such as:

height="255" width="432"/></p>

I need to strip the p tags out of resulting output.

Okay. Quibbles: translated into XSL-speak, what you
want to do is process the children of p, but not create any node for the p
itself. (XSLT doesn't know from 'tags', it knows about nodes, meaning
things like elements and attributes.) Also: you seem to be giving us
something different from <Media>....

However, if it does contain other nodes, such as:

height="255" width="432"/>This is my news release.


authors will be typing the news release content in
here! I am not sure what this news release is even
about, but lets see how it comes out in XML, shall


As I see it coming out as:</p>

I need to leave it alone.

Could your problem be restated fairly as "if there's
any text content in the paragraph, I need a paragraph in my output to
wrap it, otherwise not"?

If so, you could do:

<xsl:template match="p">
    <xsl:when test="normalize-space(.)">
      <!-- this tests true if the paragraph has any
string value
           at all after whitespace normalization.
This only occurs
           if the paragraph or any of its
descendants has non-whitespace
           text content -->

This could also be expressed as two simpler

<xsl:template match="p">

<xsl:template match="p[not(normalize-space(.))]">

Here, the first template serves as a default, and
the second one fires whenever it matches but not otherwise (it has a
higher priority since its match is more complex).

If your logic is more complex than what I'm
inferring (e.g. if it is actually more dependent on the actual type of nodes
you have down there, Media or whatever) you need to do something closer
to what you already have. Refine the logic of your requirement and you
can refine the code....


Here is what I have so far. It only looks to see if


media element or a media element wrapped by a link
element exists, but does not consider if there is a
text node after a media or link element.

<xsl:template match="p">

test="((descendant::*[1])[self::Link] and

(descendant::*[2])[self::Media]) or



Wendell Piez

XSL-List info and archive:

Current Thread