RE: [xsl] Cannot select a node here: the context item is an atomic value

["Michael Kay" <mike@xxxxxxxxxxxx>]

You'll need to set a variable outside the outer loop, and in the inner loop
do something like

<xsl:for-each select="$s/my/xpath/query[@day=$day]">

In fact, you might as well select the whole path in the variable and just
use

<xsl:for-each select="$s[@day=$day]"> 

I would tend to use .-1 rather than position()-1, but the effect is the
same. Actually, I would probably not build the duration from a string, but
rather from:

<xsl:variable name="one-day" select="xs:dayTimeDuration('P1D')"/>  (global
variable)

then in the loop:
<xsl:variable name="day" select=". * $one-day"/>

Michael Kay
http://www.saxonica.com/

> -----Original Message-----
> From: chris [mailto:oneskiingfool@xxxxxxxxx] 
> Sent: 05 April 2005 18:11
> To: xslt
> Subject: [xsl] Cannot select a node here: the context item is 
> an atomic value
> 
> Hello,
> I'm using XSLT/XPath 2.0 and am stuck on a very basic problem again. 
> All I want to do is nest my for-each loop inside another loop so that
> the inner for-each loops executes x number of times.
> To be more specific, I'm creating an HTML table with dates running
> down the left column and corresponding values in the rest of the
> cells.
> So I have date range: $date1 to $date2
> I calculated the number of days between the two: $dayCount
> 
> <xsl:for-each select="1 to $dayCount">
>    <xsl:variable name="day" select="($date1+
> xdt:dayTimeDuration(concat('P',position()-1,'D')))"/>
>    <tr>
>      <td>
>        <xsl:value-of 
> select="format-dateTime($day,'[Y,*-2]-[M,2]-[D,2]')"/>
>      </td>
>      <xsl:for-each select="my/xpath/query[@day=$day]">
>        <td>
>           <xsl:value-of select="."/>
>        </td>
>      </xsl:for-each>
>    </tr>
> </xsl:for-each>
> 
> 
> The inner loop gives me this error:
> Cannot select a node here: the context item is an atomic value
> 
> Could somebody kindly let me know how to do this?  It's clearly stated
> in the documentation that I can't do it this way (nest node loops in
> atomic loops), but I don't see a workaround.
> 
> P.S. Also any hints if there is a better way to do that outer loop
> would be appreciated as well.
> 
> Thanks!