Question

I need to transform an XML file by adding a new element that will have a value based on data in the current file and two other XML files using XSLT 1.0. The files:

File1:

<Table>
    <Row>
        <ColA>AValue1</ColA>
        <ColB>BValue1</ColB>
    </Row>
    <Row>
        <ColA>AValue2</ColA>
        <ColB>BValue2</ColB>
    </Row>
</Table>

File2:

<Table>
    <Row>
        <ColA>AValue1</ColA>
        <ColB>BValue1</ColB>
        <ColC>CValue1</ColC>
    </Row>
    <Row>
        <ColA>AValue1</ColA>
        <ColB>BValue1</ColB>
        <ColC>CValue1</ColC>
    </Row>
<Row>
        <ColA>AValue1</ColA>
        <ColB>BValue1</ColB>
        <ColC>CValue2</ColC>
    </Row>
    <Row>
        <ColA>AValue1</ColA>
        <ColB>BValue1</ColB>
        <ColC>CValue3</ColC>
    </Row>
    <Row>
        <ColA>AValue2</ColA>
        <ColB>BValue2</ColB>
        <ColC>CValue1</ColC>
    </Row>
</Table>

File3:

<Table>
    <Row>
        <ColC>CValue1</ColC>
        <ColD>ABC</ColD>
    </Row>
    <Row>
        <ColC>CValue2</ColC>
        <ColD>DEF</ColD>
    </Row>
    <Row>
        <ColC>CValue3</ColC>
        <ColD>DEF</ColD>
    </Row>
</Table>

Rows in File1 have a one-to-many relationship with rows in File2 by ColA and ColB.

Rows in File2 have a many-to-one relationship with rows in File3 by ColC.

For each row in File1, I need to:

1. Look up distinct ColC values in File2 for rows matching on ColA and ColB

2. For each distinct ColC value, look up ColD value in File3 for rows matching on ColC

3. Count the number of occurences of ColD values looked up in step 2. ColD will have one of two values (say “ABC” or “DEF”). I need to know if there are more “ABC” than “DEF” and if so add <ColD>ABC</ColD> to that Row in File1. Otherwise add <ColD>DEF</ColD> to that Row in File1. At the end, each Row in File1 should have <ColD>ABC</ColD> or <ColD>DEF</ColD>.

Desired Result (File1 transformed):

<Table>
    <Row>
        <ColA>AValue1</ColA>
        <ColB>BValue1</ColB>
        <ColD>DEF</ColD>
    </Row>
    <Row>
        <ColA>AValue2</ColA>
        <ColB>BValue2</ColB>
        <ColD>ABC</ColD>
    </Row>
</Table>

<ColD>DEF</ColD> would be added to the first Row since there were two occurences of DEF compared to 1 (distinct) occurence of ABC. <ColD>ABC</ColD> would be added to Row 2 since there was one occurence of ABC and zero DEF.

Answer 1

I came up with a solution that appears to work (haven’t tested it very much). I’m retyping it here so beware of typos. The thing I had the hardest time with was figuring out how to select unique ColC values. Looking at the solution again I didn’t need unique values. I had it in my mind that I would loop thru $unique_c_values to count ColD values, but it turned out differently.

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

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

        <xsl:variable name="unique_c_values" select=
            "document('file2.xml')/Table/Row[current()/ColA = ColA and current()/ColB = ColB]/ColC[not(. = preceding-sibling::*/ColC[current()/ColA = ../ColA and current()/ColB = ../ColB])]"/>

        <xsl:variable name="d_values" select=
            "document('file3.xml')/Table/Row[ColC = $unique_c_values]/ColD"/>

        <xsl:choose>
            <xsl:when test="count($d_values[. = 'DEF']) &gt; count($d_values[. = 'ABC'])">
                <ColD>DEF</ColD>
            </xsl:when>
            <xsl:otherwise>
                <ColD>ABC</ColD>
            </xsl:otherwise>
        </xsl:choose>
    </xsl:copy>
</xsl:template>