Re: [xsl] Handling position in non-atomic sequences

Subject: Re: [xsl] Handling position in non-atomic sequences
From: "Michael Kay mike@xxxxxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx>
Date: Sat, 2 May 2020 11:43:34 -0000
> On 2 May 2020, at 11:52, Michael MC<ller-Hillebrand mmh@xxxxxxxxx
<xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
>
> Hi folks,
>
> if a sequence stored in a variable consists of elements, we can not use the
nice function index-of().

You can, but it might not do what you want...

Saxon (PE+, 10.0+) has an extension function

saxon:index-where($seq, $predicate)

so you can do, for example, saxon:index-where($nodes, function($n){$n is
$required-node})

>
> Also, if the variable is typed as="element()*" I cannot use the axes
preceding-sibling or following-sibling.

Again, you can, but it might not do what you want. Presumably you want to find
adjacent members of the sequence. Of course a node might appear three times in
the same sequence, so just knowing the node and the sequence isn't enough...

> I can solve that by not typing the variable or use  as="document-node()"
after adding <xsl:document> to the variable's sequence constructor.

That will copy all the nodes and change their siblings and ancestry..
>
> My current solution finding the position of the first <h1> in a sequence
would be something like:

You're talking about a solution, but I don't think you've told us what the
problem is. What are you trying to achieve?
>
> count($seq/h1[1]/preceding-sibling::*) + 1
>
> with $seq being a sequence typed as document-node().
>
> Are there better methods to get the position of a certain element in the
sequence?
>

for $i in (1 to count($nodes)) return if ($nodes[$i] is $required-node) then
$i else ()

or more concisely

(1 to count($nodes))[subsequence($nodes, . 1) is $required-node]

You could also implement the saxon:index-where() function as

(1 to count($seq))[$predicate(subsequence($nodes, ., 1))]

(subsequence($nodes, $x, 1) is used rather than $nodes[$x] because predicates
change the context item)

Michael Kay
Saxonica

Current Thread