XSLT grouping by a trigger value

Question

I want to group a XML by a trigger or flag value. I can only use XSLT 1.0. Original XML:

<?xml version = "1.0" encoding = "utf-8"?>
<root>
    <image> 
        <field   level = "system" name = "Image Filename" value = "Batch121_2_1.tif"/>
        <field   level = "document" name = "groupId" value = "1"/>
        <field   level = "document" name = "scanDokumentPos" value = "1 "/>
    </image> 
    <image> 
        <field   level = "system" name = "Image Filename" value = "Batch121_2_2.tif"/>
        <field   level = "document" name = "groupId" value = "0"/>
        <field   level = "document" name = "scanDokumentPos" value = "1 "/>
    </image> 
    <image> 
        <field   level = "system" name = "Image Filename" value = "Batch121_2_3.tif"/>
        <field   level = "document" name = "groupId" value = "0"/>
        <field   level = "document" name = "scanDokumentPos" value = "1 "/>
    </image> 
    <image> 
        <field   level = "system" name = "Image Filename" value = "Batch121_3_1.tif"/>
        <field   level = "document" name = "groupId" value = "0"/>
        <field   level = "document" name = "scanDokumentPos" value = "2 "/>
    </image> 
    <image> 
        <field   level = "system" name = "Image Filename" value = "Batch121_3_2.tif"/>
        <field   level = "document" name = "groupId" value = "0"/>
        <field   level = "document" name = "scanDokumentPos" value = "2 "/>
    </image> 
    <image> 
        <field   level = "system" name = "Image Filename" value = "Batch121_7_1.tif"/>
        <field   level = "document" name = "groupId" value = "1"/>
        <field   level = "document" name = "scanDokumentPos" value = "6 "/>
    </image> 
    <image> 
        <field   level = "system" name = "Image Filename" value = "Batch121_7_2.tif"/>
        <field   level = "document" name = "groupId" value = "0"/>
        <field   level = "document" name = "scanDokumentPos" value = "6 "/>
    </image> 
    <image> 
        <field   level = "system" name = "Image Filename" value = "Batch121_8_1.tif"/>
        <field   level = "document" name = "groupId" value = "0"/>
        <field   level = "document" name = "scanDokumentPos" value = "7 "/>
    </image> 
    <image> 
        <field   level = "system" name = "Image Filename" value = "Batch121_8_2.tif"/>
        <field   level = "document" name = "groupId" value = "0"/>
        <field   level = "document" name = "scanDokumentPos" value = "7 "/>
    </image>
    <image> 
        <field   level = "system" name = "Image Filename" value = "Batch121_9_1.tif"/>
        <field   level = "document" name = "groupId" value = "1"/>
        <field   level = "document" name = "scanDokumentPos" value = "8 "/>
    </image> 
    <image> 
        <field   level = "system" name = "Image Filename" value = "Batch121_10_1.tif"/>
        <field   level = "document" name = "groupId" value = "1"/>
        <field   level = "document" name = "scanDokumentPos" value = "9 "/>
    </image> 
</root>

Result should be:

<document>
    <childdocuments>
        <document GroupID=""> <!-- grouped because of groupId=1 -->
            <childdocuments>
                <document GroupID="1"> <!-- scanDokumentPos --> <!-- grouped because of groupId=0 and scanDokumentPos=1 -->
                    <pages> 
                        <page path="Batch121_2_1.tif"> <!-- Image Filename -->
                        <page path="Batch121_2_2.tif"> <!-- Image Filename -->
                        <page path="Batch121_2_3.tif"> <!-- Image Filename -->
                    </page>
                </document>
                <document GroupID="2"> <!-- scanDokumentPos --> <!-- grouped because of groupId=0 and scanDokumentPos=2 -->
                    <pages> 
                        <page path="Batch121_3_1.tif"> <!-- Image Filename -->
                        <page path="Batch121_3_2.tif"> <!-- Image Filename -->
                    </page>
                </document>             
            </childdocuments>
        </document>
        <document GroupID=""> <!-- start new document because of groupId=1 -->
            <childdocuments>
                <document GroupID="6"> <!-- scanDokumentPos --> <!-- grouped because of groupId=0 and scanDokumentPos=6 -->
                    <pages> 
                        <page path="Batch121_7_1.tif"> <!-- Image Filename -->
                        <page path="Batch121_7_2.tif"> <!-- Image Filename -->
                    </page>
                </document>
                <document GroupID="7"> <!-- scanDokumentPos --> <!-- grouped because of groupId=0 and scanDokumentPos=7 -->
                    <pages> 
                        <page path="Batch121_8_1.tif"> <!-- Image Filename -->
                        <page path="Batch121_8_2.tif"> <!-- Image Filename -->
                    </page>
                </document>     
            </childdocuments>
        </document>     
        <document GroupID=""> <!-- start new document because of groupId=1 -->
            <childdocuments>
                <document GroupID="8"> <!-- scanDokumentPos --> <!-- grouped because of groupId=0 and scanDokumentPos=8 -->
                    <pages> 
                        <page path="Batch121_9_1.tif"> <!-- Image Filename -->
                    </page>
                </document>     
            </childdocuments>
        </document>     
        <document GroupID=""> <!-- start new document because of groupId=1 -->
            <childdocuments>
                <document GroupID="9"> <!-- scanDokumentPos --> <!-- grouped because of groupId=0 and scanDokumentPos=9 -->
                    <pages> 
                        <page path="Batch121_10_1.tif"> <!-- Image Filename -->
                    </page>
                </document>     
            </childdocuments>
        </document>     
    </childdocuments>
</document>

The first grouping-key is groupId. If 1 start new document on level 1. So all elements after groupId=1 until next groupId=1 or the end of file belongs to one group. The second grouping-key is scanDokumentPos. Group all the pages to one document which has the same scanDokumentPos.

Testing: http://xsltransform.net/3N7GxDx/3

Where is your attempted XSLT? Testing link only shows default. Please [research](https://meta.stackoverflow.com/questions/261592/how-much-research-effort-is-expected-of-stack-overflow-users) for an answer and come back with a *specific* question on earnest attempt. This grouping need in XSLT 1.0 is pretty extensive for us volunteering our free time. — Parfait, Sep 02 '20 at 15:19
Grouping in XSLT 1.0 is best done using the Muenchian method: http://www.jenitennison.com/xslt/grouping/muenchian.html For an example of grouping and subgrouping see:https://stackoverflow.com/a/58525214/3016153 — michael.hor257k, Sep 02 '20 at 15:57

Daniel Haley · Accepted Answer · 2020-09-04T16:29:12.003

Like mentioned in the comments, grouping in XSLT 1.0 is best done using Muenchian Grouping.

The example given (https://stackoverflow.com/a/58525214/3016153) is a good one, but the difference with your question is your first grouping is more like a "group-starting-with" than just grouping on a value.

What I would do is create 2 keys.

The first selects image elements that don't have a groupId of "1" based on the generated id of the first preceding sibling that has a groupId of "1".

The second selects image elements that don't have a groupId of "1" based on a combination of the generated id (see above) and the "scanDokumentPos" value.

Here's an example. I didn't handle the processing of the image elements. I only show the grouping. The handling of the image elements should be trivial.

<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
  <xsl:output indent="yes"/>
  <xsl:strip-space elements="*"/>
  
  <xsl:key name="doc_group" match="image[not(field[@name='groupId']/@value='1')]" 
    use="generate-id(preceding-sibling::image[field[@name='groupId']/@value='1'][1])"/>
  
  <xsl:key name="doc_pos_group" match="image[not(field[@name='groupId']/@value='1')]" 
    use="concat(generate-id(preceding-sibling::image[field[@name='groupId']/@value='1'][1]),'|',
    field[@name='scanDokumentPos']/@value)"/>
  
  <xsl:template match="@*|node()">
    <xsl:copy>
      <xsl:apply-templates select="@*|node()"/>
    </xsl:copy>
  </xsl:template>
  
  <xsl:template match="/*">
    <document>
      <childdocuments>
        <xsl:for-each select="image[field[@name='groupId']/@value='1']">
          <xsl:variable name="curr_image" select="."/>
          <xsl:variable name="curr_id" select="generate-id()"/>
          <document GroupID="">
            <childdocuments>
              <xsl:if test="1 >= count(key('doc_group',$curr_id))">
                <document GroupID="{normalize-space(field[@name='scanDokumentPos']/@value)}">
                  <pages>
                    <xsl:apply-templates select="$curr_image"/>
                  </pages>
                </document>
              </xsl:if>
              <xsl:for-each select="key('doc_group',$curr_id)[count(.|key('doc_pos_group', 
                concat($curr_id,'|',field[@name='scanDokumentPos']/@value))[1])=1]">
                <xsl:variable name="doc_pos" select="field[@name='scanDokumentPos']/@value"/>
                <document GroupID="{normalize-space($doc_pos)}">
                  <pages>
                    <xsl:apply-templates 
                      select="$curr_image[field[@name='scanDokumentPos']/@value=$doc_pos]|
                      key('doc_pos_group',concat($curr_id,'|',$doc_pos))"/>
                  </pages>
                </document>
              </xsl:for-each>
            </childdocuments>
          </document>
        </xsl:for-each>
      </childdocuments>
    </document>
  </xsl:template>
  
</xsl:stylesheet>

Fiddle: http://xsltfiddle.liberty-development.net/a9HjZV/4

Thanks for your solution. I noticed that my example was not so clear. It is also possible that there are several images coming, which have 1. I have adjusted the question. — knobli, Sep 04 '20 at 14:59
@knobli - Please see my update. I added an if statement that checks the size of the group. There's probably a better way to account for the single images, but unfortunately I don't have time to spend on this today. Hopefully the change helps! — Daniel Haley, Sep 04 '20 at 15:24
I tried to apply this procedure to a second case, but somehow I get 'document' as a node in the template instead of the fields. I think that I only get 1 document in the first childDocuments has the same mistake: http://xsltfiddle.liberty-development.net/a9HjZV/2 — knobli, Sep 04 '20 at 15:57
@knobli - I think the test in the "if" needs an adjustment. Does this look better? http://xsltfiddle.liberty-development.net/a9HjZV/3 If so I'll update my answer. — Daniel Haley, Sep 04 '20 at 16:20
Went ahead and updated my answer since nothing changes in the output of the fiddle. — Daniel Haley, Sep 04 '20 at 16:29
Perfect, thanks a lot. Here the final result: http://xsltfiddle.liberty-development.net/a9HjZV/5 — knobli, Sep 05 '20 at 09:36

XSLT grouping by a trigger value

1 Answers1