问题
I need to merge the values of multiple elements and take the average of the attribute field.
[INPUT]
<xml>
<characters>
<char a="a1" b="b1" y="y1" z="z1" start="1" weight="100">F</char>
<char a="a2" b="b2" y="y2" z="z2" start="0" weight="80">r</char>
<char a="a3" b="b3" y="y3" z="z3" start="0" weight="80">o</char>
<char a="a4" b="b4" y="y4" z="z4" start="0" weight="100">m</char>
<char a="a5" b="b5" y="y5" z="z5"> </char>
<char a="a6" b="b6" y="y6" z="z6" start="1" weight="100">a</char>
<char a="a7" b="b7" y="y7" z="z7" start="0" weight="80">n</char>
<char a="a8" b="b8" y="y8" z="z8" start="0" weight="80">d</char>
<char a="a9" b="b9" y="y9" z="z9"> </char>
</characters>
</xml>
[OUTPUT]
<xml>
<data>
<word>
<value>From</value>
<coordinates>a1 b1 y4 z4</coordinates>
<avgconfidence>90</avgconfidence>
</word>
<word>
<value>and</value>
<coordinates>a6 b6 y9 z9</coordinates>
<avgconfidence>90</avgconfidence>
</word>
<data>
</xml>
I have written a partial XSL below by taking suggestion of @michael.hor257k in older thread . I did try with xsl:choose. it did not yield the expected result
[XSLT]
<xsl:stylesheet version="1.0"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
<xsl:output method="xml" omit-xml-declaration="yes" indent="yes" />
<xsl:template match="/">
<data>
<xsl:for-each select="characters">
<word>
<value>
<xsl:for-each select="char">
<xsl:variable name="rec" select="." />
<xsl:if test="$rec/@start='1'">
<xsl:text> <xsl:value-of select="$rec" /></xsl:text>
</xsl:if>
</xsl:for-each>
</value>
<coordinates>
<xsl:value-of select="char[1]/@a"/>
<xsl:text> </xsl:text>
<xsl:value-of select="char[1]/@b"/>
<xsl:text> </xsl:text>
<xsl:value-of select="char[last()]/@y"/>
<xsl:text> </xsl:text>
<xsl:value-of select="char[last()]/@z"/>
</coordinates>
<avgWeight>
<xsl:value-of select="sum(characters/char/@weight) div count(characters/char) "/>
</avgWeight>
</word>
</xsl:for-each>
</data>
</xsl:template>
</xsl:stylesheet>
Please help.
回答1:
Interesting problem - your partial XSL will work for a single word but you need a way to segment your characters into separate words and handle each word separately. I see it as a grouping task, where a "word" consists of a char[@start='1']
element along with its following char[@start='0']
elements. So you could approach it using a key to associate each non-word-initial char
with the start of its word (i.e. its nearest preceding start='1'
sibling)
<xsl:stylesheet version="1.0"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
<xsl:output method="xml" omit-xml-declaration="yes" indent="yes" />
<xsl:key name="wordChars" match="char[@start='0']"
use="generate-id(preceding-sibling::char[@start='1'][1])" />
<xsl:template match="/">
<xml>
<data>
<xsl:apply-templates select="xml/characters/char[@start='1']" />
</data>
</xml>
</xsl:template>
<xsl:template match="char">
<!-- all the chars in this word - . is the initial letter, the key gives the
rest of the word -->
<xsl:variable name="word" select=". | key('wordChars', generate-id())" />
<word>
<value>
<xsl:for-each select="$word"><xsl:value-of select="."/></xsl:for-each>
</value>
<coordinates>
<xsl:value-of select="concat(
$word[1]/@a, ' ', $word[1]/@b, ' ',
$word[last()]/@y, ' ', $word[last()]/@z)" />
</coordinates>
<avgconfidence>
<xsl:value-of select="sum($word/@weight) div count($word)" />
</avgconfidence>
</word>
</xsl:template>
</xsl:stylesheet>
When run over your sample input this produces
<xml>
<data>
<word>
<value>From</value>
<coordinates>a1 b1 y4 z4</coordinates>
<avgconfidence>90</avgconfidence>
</word>
<word>
<value>and</value>
<coordinates>a6 b6 y8 z8</coordinates>
<avgconfidence>86.6666666666667</avgconfidence>
</word>
</data>
</xml>
which I believe to be the correct answer (with end co-ordinates y8 z8
rather than y9 z9
and confidence 862/3 rather than 90 for the second word).
来源:https://stackoverflow.com/questions/27843380/merge-the-values-of-multiple-elements-and-take-the-average-of-the-attribute-fiel