Looping through nodes in XML using XSLT

Im not sure, but if you are looking for children of a particular node, eg data

you could use : for each instance of data look for children (*) which have @value

so

data/*/@value

should be the right xpath

MM

I need to be able to be more specific with the selection, e.g.

<xsl:for-each select="T[*]">
   <xsl:value-of select="T[*]/@value"/>
</xsl:for-each>

Or something like that (I assume that doesn't work)

<xsl:for-each select="T[*]">
   <xsl:value-of select="T[*]/@value"/>
</xsl:for-each>

my previous post should have given you that
please note that inside the for-each loop,
the context changes
so    <xsl:value-of select="T[*]/@value"/>
will not work
   <xsl:value-of select="@value"/>
will work

Hi Gertone,

Thanks for that - although my response came after yours I hadn't seen your post. Your post looks promising ... I'll give it a go. Thanks.

webtechy,

To take away all confusion
select="T[*]"
is the element with name 'T' that has a subelement (name '*', so undefined)

before the [] you have the path of the node you are looking for
the [] (called predicate) creates a specific condition

what you wrote in pseudocode
is *[starts-with(name(), 'T')] in real XPath

the element '*' (means can be any name)
that has a name that starts with a 'T'

there is a way to guarantuee that the remainder is an integer but that is a bit more complex

If you are sure you don't have an element like "Tate" you don't won't to address,
I would leave it like this

cheers

webtechy,
>  I hadn't seen your post.

I already guessed that, they came in the same minute :-)

webtechy,
> <Data>
>   <Date value="14 Nov 2005"/>
>   <Location value="1001"/>
>   <InStation1 value="Y"/>
>   <T1 value="A"/>
>   <InStation2 value="N"/>
>   <T2 value="B"/>
>   <InStation3 value="Y"/>
>   <T3 value="B"/>
> </Data>

as you said, you can't change the XML, you probably realise this is very bad modelling practice.
With the element names you should try to give some sort of meaningfull structure
dynamic data belongs in an attribute value or the content, not in ever changing element names.
XML created this way, is very hard to process or validate
or very difficult to document

On top of that I have the feeling this XML tries to combine values based on the number in the tagname

I would definitely go for an XML like this

<Data>
  <Date value="14 Nov 2005"/>
  <Location value="1001"/>
  <Station id="1">
    <InStation>Y</InStation>
    <T>A</T>
  </Station>
  <Station id="2">
    <InStation>N</InStation>
    <T>B</T>
  </Station>
  <Station id="3">
    <InStation>Y</InStation>
    <T>B</T>
  </Station>
 </Data>

This gives you an explicit grouping of "Instation" and "T"
and a clear model of Station

but perhaps you already knew
cheers

Geert