0

I'm trying to transform old XML files into a new and improved structure. Part of this needs me to wrap some loose elements into a parent container, as well as modify their children

Old structure

<monograph>
  <title>asdf</title>

  <dosage.sec id="dosage.sec.1">
    <dosage.sec>asgfd</dosage.sec>
    <dosage.sec>asgfd</dosage.sec>
  </dosage.sec>
  <dosage.sec id="dosage.sec.2">
    <dosage.sec>asgfd</dosage.sec>
    <dosage.sec>asgfd</dosage.sec>
  </dosage.sec>
  <dosage.sec id="dosage.sec.3">
    <dosage.sec>asgfd</dosage.sec>
    <dosage.sec>asgfd</dosage.sec>
  </dosage.sec>

  <products>
    <prod>sadf</prod>
    <prod>sadf</prod>
  </products>
</monograph>

New structure

<monograph>
  <title>asdf</title>

  <dosage>
    <dosage.sec id="dosage.sec.1">
      <dosage.qual>asgfd</dosage.qual>
      <dosage.qual>asgfd</dosage.qual>
    </dosage.sec>
    <dosage.sec id="dosage.sec.2">
      <dosage.qual>asgfd</dosage.qual>
      <dosage.qual>asgfd</dosage.qual>
    </dosage.sec>
    <dosage.sec id="dosage.sec.3">
      <dosage.qual>asgfd</dosage.qual>
      <dosage.qual>asgfd</dosage.qual>
    </dosage.sec>
  </dosage>

  <products>
    <prod>sadf</prod>
    <prod>sadf</prod>
  </products>
</monograph>

I found this answer and modified it a bit to suit my needs:

<!-- wrap dosage.sec elements in a dosage container -->
<xsl:template match="node()|@*" name="dosage.sec">
    <xsl:copy>
        <xsl:apply-templates select="node()|@*" />
    </xsl:copy>
</xsl:template>

<xsl:template match="monograph/dosage.sec[not(preceding-sibling::*[1][self::dosage.sec])]">
    <dosage>
        <xsl:call-template name="dosage.sec" />
        <xsl:apply-templates mode="copy" select="following-sibling::*[1][self::dosage.sec]" />
    </dosage>
</xsl:template>

<xsl:template match="monograph/dosage.sec" mode="copy">
    <xsl:call-template name="dosage.sec"/>
</xsl:template>


<!-- rename children dosage.sec -->
<xsl:template match="dosage.sec/dosage.sec">
    <dosage.qual>
        <xsl:apply-templates />
    </dosage.qual>
</xsl:template>

But my output is:

<monograph>
  <title>asdf</title>

  <dosage>
    <dosage.sec id="dosage.sec.1">
      <dosage.qual>asgfd</dosage.qual>
      <dosage.qual>asgfd</dosage.qual>
    </dosage.sec>
    <dosage.sec id="dosage.sec.2">
      <dosage.qual>asgfd</dosage.qual>
      <dosage.qual>asgfd</dosage.qual>
    </dosage.sec>
    <dosage.sec id="dosage.sec.3">
      <dosage.qual>asgfd</dosage.qual>
      <dosage.qual>asgfd</dosage.qual>
    </dosage.sec>
  </dosage>

  <dosage.sec id="dosage.sec.3">
    <dosage.qual>asgfd</dosage.qual>
    <dosage.qual>asgfd</dosage.qual>
  </dosage.sec>

  <products>
    <prod>sadf</prod>
    <prod>sadf</prod>
  </products>
</monograph>

I'm using PHP5's built-in XSLTProcessor object - all XML and XSL are version 1.0

Community
  • 1
  • 1
HorusKol
  • 8,375
  • 10
  • 51
  • 92
  • Your XSLT would probably work if you removed the `[1]` from this line: `` but the way you're going about it may not be the best approach. – JLRishe Mar 05 '13 at 10:34
  • The answer you referred to was addressing a particular scenario where they needed to wrap items that had different items before and after them. Do you have a requirement like that, or is it as straightforward as your example? – JLRishe Mar 05 '13 at 16:19
  • All of the dosage.sec elements should be contiguous within the source document (no other elements in between), so it is simpler. I've modified my example, as I realised that I didn't accurately depict that I have other elements in the monograph before and after the block of dosage.sec – HorusKol Mar 05 '13 at 22:58

2 Answers2

2

How about this:

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

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

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

  <!-- rename children dosage.sec -->
  <xsl:template match="dosage.sec/dosage.sec">
    <dosage.qual>
      <xsl:apply-templates />
    </dosage.qual>
  </xsl:template>
</xsl:stylesheet>

When run on your sample input, this produces:

<monograph>
  <dosage>
    <dosage.sec>
      <dosage.qual>asgfd</dosage.qual>
      <dosage.qual>asgfd</dosage.qual>
    </dosage.sec>
    <dosage.sec>
      <dosage.qual>asgfd</dosage.qual>
      <dosage.qual>asgfd</dosage.qual>
    </dosage.sec>
    <dosage.sec>
      <dosage.qual>asgfd</dosage.qual>
      <dosage.qual>asgfd</dosage.qual>
    </dosage.sec>
  </dosage>
</monograph>
JLRishe
  • 99,490
  • 19
  • 131
  • 169
0

I had a few other problems on the way - for one, I discovered that the dosage.sec elements were not in contiguous block and I was erroneously creating two dosage containers.

These templates solved my problems:

<!-- wrap dosage.sec elements in a dosage container -->
<xsl:template match="monograph/dosage.sec" name="dosage.sec">
    <xsl:param name="drugname" />

    <xsl:copy>
        <xsl:apply-templates>
            <xsl:with-param name="drugname" select="$drugname" />
        </xsl:apply-templates>
    </xsl:copy>
</xsl:template>

<xsl:template match="monograph/dosage.sec[not(preceding::dosage.sec)]">
    <xsl:param name="drugname" />

    <dosage>
        <xsl:call-template name="dosage.sec">
            <xsl:with-param name="drugname" select="$drugname" />
        </xsl:call-template>

        <xsl:apply-templates mode="copy" select="following::dosage.sec">
            <xsl:with-param name="drugname" select="$drugname" />
        </xsl:apply-templates>
    </dosage>
</xsl:template>

<xsl:template match="monograph/dosage.sec" mode="copy">
    <xsl:param name="drugname" />

    <xsl:call-template name="dosage.sec">
        <xsl:with-param name="drugname" select="$drugname" />
    </xsl:call-template>
</xsl:template>

<xsl:template match="monograph/dosage.sec[(preceding::dosage.sec)]" />
HorusKol
  • 8,375
  • 10
  • 51
  • 92