Template is called where it should not to be - xslt

Question

I've the below two XML cases.
Case1:
<para>Rent is the sum of money paid by the Tenant to the Landlord for the exclusive use of premises. The Landlord and Tenant signs a <page num="4"/>tenancy agreement which has to be stamped with the tax authorities as required under the Stamp Duties Act. The stamping of a tenancy agreement gives it validity but if the tenancy agreement is not stamped that does not mean</para>
Case2:
<para><page num="5"/>The Writ of Distress proceedings is an effective way to recover arrears in rent but regard must be had to the Landlord/Tenant relationship and the effect of publicity of such proceedings to the image of the building amongst other things.</para>
and the below XSLT
<xsl:template match="para">
<xsl:apply-templates select="child::node()[(self::page)]"/>
<li class="item">
<div class="para">
<span class="item-num">
<xsl:value-of select="../#num"></xsl:value-of>
</span>
<xsl:apply-templates select="child::node()[not(self::page)]"/>
</div>
</li>
</xsl:template>
<xsl:template match="page">
<xsl:processing-instruction name="pb">
<xsl:text>label='</xsl:text>
<xsl:value-of select="./#num"/>
<xsl:text>'</xsl:text>
<xsl:text>?</xsl:text>
</xsl:processing-instruction>
<a name="{concat('pg_',./#num)}"/>
<xsl:apply-templates/>
</xsl:template>
What I'm trying to do is check if page is the immediate(first) child of para and print that value first and then do the rest. But in both the cases, the page is printed first.
In the above cases provided, for case1, the page should be called just like any other template in para, since it is not the immediate child of para, but in case2, first the page has to be printed and next the template is to be called, as page num="5" is the immediate child of para Please let me know how I can do this.
A demo is here

Answer 1

I think what you mean is that you want to perform extra processing when page is the first child node under para. Your apply-templates need to look like this
<xsl:apply-templates select="node()[1][self::page]" />
However, it also sounds like you want to perform other processing on page elements regardless. You probably need two templates matching page here, but one with a "mode" to distinguish it from your normal processing.
Call it like this
<xsl:apply-templates select="node()[1][self::page]" mode="first"/>
And match it like this
<xsl:template match="page" mode="first">
This would contain the code to output your processing instruction.
For "normal" processing of the page element, just have another template matching page without the mode
<xsl:template match="page">
<a name="{concat('pg_',./#num)}"/>
<xsl:apply-templates/>
</xsl:template>
Try this XSLT
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="2.0">
<xsl:output indent="yes"/>
<xsl:strip-space elements="*" />
<xsl:template match="#*|node()">
<xsl:copy>
<xsl:apply-templates select="#*|node()"/>
</xsl:copy>
</xsl:template>
<xsl:template match="para">
<xsl:apply-templates select="node()[1][self::page]" mode="first"/>
<li class="item">
<div class="para">
<span class="item-num">
<xsl:value-of select="../#num"></xsl:value-of>
</span>
<xsl:apply-templates />
</div>
</li>
</xsl:template>
<xsl:template match="page">
<a name="{concat('pg_',./#num)}"/>
<xsl:apply-templates/>
</xsl:template>
<xsl:template match="page" mode="first">
<xsl:processing-instruction name="pb">
<xsl:text>label='</xsl:text>
<xsl:value-of select="./#num"/>
<xsl:text>'</xsl:text>
</xsl:processing-instruction>
</xsl:template>
</xsl:stylesheet>
EDIT: If you don't want both "page" templates to apply to apply to the first page element, then add the following template to ignore it
<xsl:template match="page[not(preceding-sibling::node())]" />
Note, this will only work if you have <xsl:strip-space elements="*" /> present in your document, to strip out white-space only text nodes. Alternatively, you could write this
<xsl:template match="page[not(preceding-sibling::node()[not(self::text()) or normalize-space()])]" />
EDIT 2: The reason you need the extra templates is because of this line
<xsl:apply-templates />
This is will look for templates that match all the child elements under the current para element. So, for a page element, the following template will be matched
<xsl:template match="page">
But you say you don't want the very first page element to be matched in this case. Therefore, you are a more 'specific' template to match it. For example
<xsl:template match="page[not(preceding-sibling::node())]" />
This template matches page elements with no preceding siblings; i.e. the very first element under para.
XSLT has the concept of priority for templates. Where a template matching an element with a condition specified, that template will always be given priority. In this case, the specific template simply ignores the page element, to ensure it doesn't get output.
For other page elements, the other template will be used as normal.