CODEWITHSUNDEEP
xmlcsvxslt
Rusty Lemur
Rusty Lemur

Reputation: 1875

Transform multi-element XML to CSV with XSLT

I am trying to understand how XSLT can iterate over multiple elements and output data from the those elements and their older and younger siblings. See the example below:

<a>
    <b name="first">
        <c value="C Value">C Text</c>
        <d>D1 Text</d>
        <d>D2 Text</d>
        <e>E Text</e>
    </b>
    <b name="second">
        <c value="C Value">C Text</c>
        <d>D1 Text</d>
        <d>D2 Text</d>
        <d>D3 Text</d>
        <e>E Text</e>
    </b>
    <b name="third">
        <c value="C Value">C Text</c>
        <e>E Text</e>
    </b>
</a>

I would like the output to be like the following (assuming for simplicity that each element's text will not contain commas).

first,C Value,C Text,D1 Text,E Text
first,C Value,C Text,D2 Text,E Text
second,C Value,C Text,D1 Text,E Text
second,C Value,C Text,D2 Text,E Text
second,C Value,C Text,D3 Text,E Text
third,C Value,C Text,,E Text

So there could be an arbitrary number of <d> elements (or none at all). Each CSV line must contain the information from <c> (the older sibling of <d>), the information from one instance of <d>, and the information from <e>, the younger sibling of <d>. (I may be making up this notion of younger and older, but it seems to make sense.)

I haven't worked with XSLT before, so I really don't know where to start with it. The examples I've found for iterating over elements do not make it clear how I can pull values from later in the document (e.g. <e> in the example) and return to another instance of the element (e.g. <d>). I've written an implementation of this in Python with lxml, but would like to see if XSLT is better suited for this transformation.

EDIT: I just saw that an answer was posted which I will be studying in detail to try to understand it. But just to share what I've been working on, here is the XSL that I have been developing. It fails to output anything when there is no <d> value.

<?xml version="1.0" encoding="utf-8"?>
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">

    <xsl:output method="text" encoding="iso-8859-1" />
    <xsl:strip-space elements="*" />

    <xsl:template match="a">       
        <xsl:for-each select="b">
            <xsl:call-template name="match-b"/>
        </xsl:for-each>
        <xsl:text>&#10;</xsl:text>
    </xsl:template>

    <xsl:template name="match-b" match="b">
        <xsl:for-each select="d">
            <xsl:value-of select="../@name"/>
            <xsl:text>,</xsl:text>
            <xsl:value-of select="../c/@value"/>
            <xsl:text>,</xsl:text>
            <xsl:value-of select="../c"/>
            <xsl:text>,</xsl:text>
            <xsl:value-of select="."/>
            <xsl:text>,</xsl:text>
            <xsl:value-of select="../e"/>
            <xsl:text>&#10;</xsl:text>
        </xsl:for-each>
    </xsl:template>

</xsl:stylesheet>

Output:

first,C value,C Text,D1 Text,E Text
first,C value,C Text,D2 Text,E Text
second,C value,C Text,D1 Text,E Text
second,C value,C Text,D2 Text,E Text
second,C value,C Text,D3 Text,E Text

Upvotes: 0

Views: 214

Answers (1)

Martin Honnen
Martin Honnen

Reputation: 167561

You can navigate to preceding or following sibling or the parent node using XPath. To output a line with XSLT 2 or later you can use xsl:value-of select="expression-to-compute-column-values" separator=",", as in

<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
    version="3.0">

  <xsl:output method="text"/>

  <xsl:mode name="add-dummy-d" on-no-match="shallow-copy"/>

  <xsl:variable name="normalized-input">
      <xsl:apply-templates mode="add-dummy-d"/>
  </xsl:variable>

  <xsl:template match="b[not(d)]/c" mode="add-dummy-d">
      <xsl:next-match/>
      <d/>
  </xsl:template>

  <xsl:template match="/">
      <xsl:apply-templates select="$normalized-input/a/b/d"/>
  </xsl:template>

  <xsl:template match="d">
      <xsl:value-of select="../@name, preceding-sibling::c/(@value, .), ., following-sibling::e" separator=","/>
      <xsl:text>&#10;</xsl:text>
  </xsl:template>

</xsl:stylesheet>

As said in a comment, an intermediary step adding an empty d for those b that don't have is needed or at least one way to ensure you get the last row of your sample output.

Online sample at https://xsltfiddle.liberty-development.net/eiZQaF7.

Upvotes: 2

Related Questions