RE: Grouping of nodes

[Larry Trammell]

Hi Dario, I am a non-expert occasional user of XSLT (there's a formula for
disaster, so be warned!), but I recently learned about STX processing and
tried the Joost implementation. XSLT and STX transformations appear very
similar, and this is not an accident. The two main differences are: 1) the
STX execution model is one pass and sequential, so it does not need to build
and manipulate a representation of the entire document in memory, which
usually means reduced memory, cache, and disk utilization; 2) because of the
limited execution model, it has only limited transformation capabilities.
However, your transformation seems to be ideally suited for sequential
processing. It might be worth a try.
 

-----Original Message-----
From: Dario Laera [mailto:[EMAIL PROTECTED]
Sent: Friday, May 30, 2008 12:21 AM
To: xalan-j-users@xml.apache.org
Subject: Grouping of nodes


Hi all,
I have xml files with thousands of sibling nodes, I want to transform  
the xml so that all this nodes gets grouped in subset of n elements. I  
have wrote an xsl file for this but it's very slow and I have big  
files: 20 minutes of cpu time for 20MB file.
Here's an example code.

--input XML--
<?xml version="1.0" encoding="UTF-8"?>
<report>
        <object>1</object>
        <object>2</object>
        <object>3</object>
        <object>4</object>
        <object>5</object>
        ...
        <object>100</object>
</report>

--output XML--
<?xml version="1.0" encoding="UTF-8"?>
<report>
        <object-collection>
                <object>1</object>
                <object>2</object>
                <object>3</object>
                <object>4</object>
                <object>5</object>
        </object-collection>
        <object-collection>
                <object>6</object>
                <object>7</object>
                <object>8</object>
                <object>9</object>
                <object>10</object>
        </object-collection>
        ...
        <object-collection>
                <object>96</object>
                <object>97</object>
                <object>98</object>
                <object>99</object>
                <object>100</object>
        </object-collection>
</report>

-- XSL --
<?xml version='1.0'?>
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform";
                version='1.0'>
<xsl:output method="xml" indent="yes"/>

<xsl:template match="*|@*|text()">
    <xsl:copy>
        <xsl:apply-templates select="@*"/>
        <xsl:apply-templates/>
    </xsl:copy>
</xsl:template>

<xsl:template match="/report">
    <report>
        <xsl:for-each select="./object">
            <xsl:if test="position() mod 5 = 1">
                <object-collection>
                    <xsl:apply-templates select="."/>
                    <xsl:variable name="count" select="position()"/>
                    <xsl:apply-templates select="../object[$count +  
1]"/>
                    <xsl:apply-templates select="../object[$count +  
2]"/>
                    <xsl:apply-templates select="../object[$count +  
3]"/>
                    <xsl:apply-templates select="../object[$count +  
4]"/>
                </object-collection>
            </xsl:if>
        </xsl:for-each>
    </report>
</xsl:template>

</xsl:stylesheet>


My solution is the above, but it's not time efficient. Is there a way  
to do it better?