Cart
[Home] [By Thread] [By Date] [Recent Entries]
Hi David,
Perhaps something like this (XSLT 2.0), taken all your requirements, if I understood them correctly, but it's a rather naive attempt, I'm sure it can be far simpler than this (I loathe the repeated logic...): <xsl:template match="/" name="main"> <xsl:for-each-group select="times/time[position() != last()]" group-starting-with="node()[@time = ../../cells/cell/(@startTime, @endTime)]"> <!-- all starting times --> <xsl:apply-templates select="current-group()[1][@time = ../../cells/cell/@startTime]" mode="start-time" /> <!-- all times 'during' an event --> <xsl:apply-templates select="current-group()[1][@time = ../../cells/cell/@startTime]/current-group()[position() > 1]" /> <!-- all times that form a gap --> <xsl:apply-templates select="current-group()[1][not(@time = ../../cells/cell/@startTime)]/current-group()" mode="gap-time" /> </xsl:for-each-group> </xsl:template> <xsl:template match="time" mode="start-time"> <start-time time="{.}" user="{../../cells/cell[@startTime = current()/@time]}" /> </xsl:template> <xsl:template match="time">
<occupied-time time="{.}" />
</xsl:template><xsl:template match="time" mode="gap-time">
<gap time="{.}" />
</xsl:template>This will output, taken your input, the following XML: <start-time time="11:00" user="Marquette"/> <occupied-time time="11:30"/> <occupied-time time="12:00"/> <occupied-time time="12:30"/> <gap time="01:00"/> <gap time="01:30"/> <start-time time="02:00" user="Birnbaum"/> <occupied-time time="02:30"/> <start-time time="03:00" user="Konsko"/> <occupied-time time="03:30"/> <start-time time="04:00" user="Frasier"/> <occupied-time time="04:30"/> Since you say that you only need the 'gap' time, you can suffice with the for-each-group and the following as its content, with or without a mode, depending on what you want: <xsl:apply-templates select="current-group()[1][not(@time = ../../cells/cell/@startTime)]/current-group()" /> Can someone, btw, elaborate on the following: Why is it that a statement like current-group()[some-predicate]/following-sibling::* will return *all* following sibling and not only the elements that are part of the current group? My guess is that current-group() retains its bond with the original tree, is that correct? To workaround this, I know that you can treat the current-group() results as a document-node when you copy it into a variable and go from there. Cheers, -- Abel Braaksma David J Birnbaum wrote:
|
