Question

I want to merge items with same id into single item. The resultant item needs to have first non-empty element from the item group. If there is no non-empty element in the item group then the element itself ignored from the item group and should not present in the resultant item.

Sample XML Input:

<?xml version="1.0" encoding="UTF-8" ?>
<shiporder orderid="orderid106">
   <orderperson id="123">orderperson107</orderperson>
   <shipto>
      <name>name108</name>
      <address>address109</address>
      <city>city110</city>
      <country>country111</country>
   </shipto>

   <!--Item Group 100-->
   <item>
      <id>100</id>
      <title>
         <first>
            <a></a>
         </first>
         <last>item100_lastTitle1</last>
      </title>
      <note></note>
      <quantity></quantity>
   </item>
   <item>
      <id>100</id>
      <title>
         <first>
            <a>a_100_2</a>
         </first>
         <last>item100_lastTitle2</last>
      </title>
      <note>note100_2</note>
      <quantity></quantity>
   </item>
   <item>
      <id>100</id>
      <title>
         <first>
            <a att1="abc" att2='cde'>a_100_3</a>
         </first>
         <last id="1" attr="2">item100_lastTitle3</last>
      </title>
      <note>note100_3</note>
      <quantity>1</quantity>
   </item>

   <!--Item Group 101-->
   <item>
      <id>101</id>
      <title>
         <first>
            <a>a_101_1</a>
         </first>
         <last></last>
      </title>
      <note>note101_1</note>
      <quantity>10</quantity>
   </item>
   <item>
      <id>101</id>
      <title>
         <first>
            <a>a_101_2</a>
         </first>
         <last>item101_lastTitle2</last>
      </title>
      <note>note101_2</note>
      <quantity>5</quantity>
   </item>

   <!--Item Group 103-->
   <item>
      <id>103</id>
      <title>
         <first>
            <a>a_103_2</a>
         </first>
         <last>item103_lastTitle2</last>
      </title>
      <note>note103_1</note>
      <quantity></quantity>
   </item>
</shiporder>

Sample XML Output:

<?xml version = '1.0' encoding = 'UTF-8'?>
<shiporder orderid="orderid106">
  <item>
    <id>100</id>
    <title>
      <first>
        <a>a_100_2</a>
      </first>
      <last>item100_lastTitle1</last>
    </title>
    <note>note100_2</note>
    <quantity>6</quantity>
  </item>

  <item>
    <id>101</id>
    <title>
      <first>
        <a>a_101_1</a>
      </first>
      <last>item101_lastTitle2</last>
    </title>
    <note>note101_1</note>
    <quantity>10</quantity>
  </item>

  <item>
    <id>103</id>
    <title>
      <first>
        <a>a_103_2</a>
      </first>
      <last>item103_lastTitle2</last>
    </title>
    <note>note103_1</note>
  </item>
</shiporder>

I tried with the following XSL code to get the above output:

<?xml version="1.0" encoding="windows-1252" ?>
<xsl:stylesheet version="2.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
   <xsl:template match="/shiporder">
      <shiporder>
         <xsl:if test="@orderid">
            <xsl:attribute name="orderid">
               <xsl:value-of select="@orderid"/>
            </xsl:attribute>
         </xsl:if>

         <!-- Grouping each item based on its ID element value-->
         <xsl:for-each-group select="item" group-by="id">
            <item-group>
               <!-- For Each group search for first non empty child element-->
               <xsl:for-each select="current-group()[id/text()][1]">
                  <xsl:copy-of select="id"/>
               </xsl:for-each>
               <title>
                  <first>
                     <!-- for 'a' element -->
                     <xsl:variable name="temp1">
                        <xsl:for-each select="current-group()/title/first[a/text()][1]">
                           <elm>
                              <xsl:copy-of select="a"/>
                           </elm>
                        </xsl:for-each>
                     </xsl:variable>
                     <xsl:copy-of select="$temp1/elm[1]/a"/>
                  </first>

                  <!-- for 'last' element -->
                  <xsl:variable name="temp2">
                     <xsl:for-each select="current-group()/title[last/text()][1]">
                        <elm>
                           <xsl:copy-of select="last"/>
                        </elm>
                     </xsl:for-each>
                  </xsl:variable>
                  <xsl:copy-of select="$temp2/elm[1]/last"/>
               </title>

               <!-- for 'note' element -->
               <xsl:for-each select="current-group()[note/text()][1]">
                  <xsl:copy-of select="note"/>
               </xsl:for-each>

               <!-- for 'quantity' element -->
               <xsl:for-each select="current-group()[quantity/text()][1]">
                  <xsl:copy-of select="quantity"/>
               </xsl:for-each>

            </item-group>
         </xsl:for-each-group>
      </shiporder>
   </xsl:template>
</xsl:stylesheet>

The code works fine But the worst part is, I couldn’t make it generalize. For each element under the item I need to write a xsl code specific to the xpath of particular element. I couldn’t even write code in terms of user-defined-templates.

I found already answered question in the title of “Recursively combine identical sibling elements in XSLT“. But I’m not able to customize it according to my requirements.

I’m having hundreds of such child or decedent element under each item and the depth of each one is arbitrary. It is very ridiculous to write xsl code corresponds to each of those child or descendants. Can any of the experts guide me to write generic xsl code (irrespective of number of elements and also its depth) or optimize the existing code?

Thanks in advance,

Kathir

Answer 1

This is actually an XSLT 1.0 solution (I don’t have an XSLT2 processor to hand), but it uses a generic ‘combine’ named template that should do what you need.

<?xml version="1.0" encoding="windows-1252" ?>
<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
  <xsl:key name="item" use="id" match="item" />

  <xsl:template match="item[generate-id() = generate-id(key('item',id)[1])]">
    <xsl:call-template name="combine">
      <xsl:with-param name="list" select="../*[id = current()/id]" />
    </xsl:call-template>
  </xsl:template>

  <xsl:template match="item" />

  <xsl:template name="combine">
    <xsl:param name="list" />
    <xsl:element name="{name($list[1])}">
      <xsl:for-each select="*[not(preceding-sibling::*[name() = name(current())])]" >
        <xsl:call-template name="combine">
          <xsl:with-param name="list" select="$list/*[name() = name(current())]" />
        </xsl:call-template>
      </xsl:for-each>
      <xsl:value-of select="$list[text()][1]/text()" />
    </xsl:element>
  </xsl:template>

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

The combine template takes a list of nodes, and combines them into one recursively, taking the first available value for each node. This template does NOT combine attributes as well, but as you don’t have any in the item XML you provided, this should be OK. The <xsl:for-each in the middle of the template uses an XSLT1 technique for picking out unique element names; it only selects those that don’t have a previous sibling with the same name. There’s probably a way of doing this using for-each-group in XSLT2, but I can’t check that here.