Reputation: 17
Using Saxon 9.7, XSLT 3.0, I'm trying to select square bracketed terms from a string of text and then remove duplicate values of the terms.
So far I have found a template which selects the substrings I want and a function that tokenizes the string and then removes duplicate values. However, I haven't been able to get the correct regex for the tokenizing of the string.
Here is my XML of the full text
<column>
<columnDerivationPrompt>Option 1: (No visit windowing)</columnDerivationPrompt>
<columnDerivationDescription>Set to collected visit name [EG.VISIT] Set to 'POST-BASELINE MINIMUM' for the new observation generated for derviation type minimum [ADEG.DTYPE] = 'MINIMUM'
Set to 'POST-BASELINE MAXIMUM' for the new observation generated for derviation type maximum [ADEG.DTYPE]= 'MAXIMUM'
</columnDerivationDescription>
<columnDerivationPrompt>Option 2: (User defined visit windows)</columnDerivationPrompt>
<columnDerivationDescription>Set to a re-defined visit range based on user-defined input, using formatting of Analysis Relative Day [ADEG.ADY] range in conjunction with Analysis Window Target [ADEG.AWTARGET] and Analysis Window Diff from Target [ADEG.AWTDIFF] to determine analysis visit.
Set to 'POST-BASELINE MINIMUM' for the new observation generated for derviation type minimum [ADEG.DTYPE] = 'MINIMUM'
Set to 'POST-BASELINE MAXIMUM' for the new observation generated for derviation type maximum [ADEG.DTYPE]= 'MAXIMUM'
</columnDerivationDescription>
</column>
The string of terms taken from the text that I need to remove duplicates from
EG.VISIT ADEG.DTYPE ADEG.DTYPE ADEG.ADY ADEG.AWTARGET ADEG.AWTDIFF ADEG.DTYPE ADEG.DTYPE
What I would like to see
EG.VISIT ADEG.DTYPE ADEG.ADY ADEG.AWTARGET ADEG.AWTDIFF
my XSLT template and function
<xsl:variable name="test">
<xsl:if test="contains($string,'[')">
<xsl:variable name="relevant-part" select="substring-before(substring-after($string,'['),']')"/>
<xsl:variable name="remainder" select="substring-after($string,']')"/>
<xsl:value-of select="$relevant-part"/>
<xsl:if test="contains($remainder,'[')">
<xsl:text disable-output-escaping="yes"> </xsl:text>
</xsl:if>
<xsl:call-template name="find-relevant-text">
<xsl:with-param name="string" select="$remainder"/>
</xsl:call-template>
</xsl:if>
</xsl:variable>
<xsl:value-of select="myfn:sortCSV($test)"/>
</xsl:template>
<xsl:function name="myfn:sortCSV" as="xs:string*">
<xsl:param name="csvString" as="xs:string"/>
<!-- Split up string and remove duplicates -->
<xsl:variable name="values" select="distinct-values(tokenize($csvString,'\W+\.\W+'))" as="xs:string*"/>
<!-- Return all elements, sorted -->
<xsl:for-each select="$values">
<xsl:sort/>
<!-- We don't return empty strings -->
<xsl:sequence select=".[.!='']"/>
</xsl:for-each>
</xsl:function>
\W+\.\W+
is the regex I have been using to identify e.g. EG.VISIT or ADEG.DTYPE. So any pattern including CC.CCCC to CCCC.CCCCCCCC (where C is a char [A-Z]).
The output I am getting is
EG.VISIT ADEG.DTYPE ADEG.DTYPE ADEG.ADY ADEG.AWTARGET ADEG.AWTDIFF ADEG.DTYPE ADEG.DTYPE
So no duplicates have been removed.
QUESTION: Can anyone see where I am going wrong with my expression or code?
Upvotes: 0
Views: 4844
Reputation: 167716
I would use analyze-string
, either with XSLT 2.0 the XSLT xsl:anyalyze-string
or with XSLT 3.0 the function of the same name, using that approach it is a one-liner:
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
xmlns:xs="http://www.w3.org/2001/XMLSchema"
xmlns:fn="http://www.w3.org/2005/xpath-functions"
xmlns:math="http://www.w3.org/2005/xpath-functions/math"
exclude-result-prefixes="xs math fn"
version="3.0">
<xsl:template match="column">
<xsl:value-of select="distinct-values(analyze-string(., '\[([A-Z]+\.[A-Z]+)\]')//fn:match/fn:group[@nr = 1])"/>
</xsl:template>
</xsl:stylesheet>
Output is EG.VISIT ADEG.DTYPE ADEG.ADY ADEG.AWTARGET ADEG.AWTDIFF
.
If you want to sort the extracted strings then use <xsl:value-of select="sort(distinct-values(analyze-string(., '\[([A-Z]+\.[A-Z]+)\]')//fn:match/fn:group[@nr = 1]))"/>
.
Upvotes: 1
Reputation: 627370
As for your regular expression, note that a \W
matches a non-word char and cannot match uppercase (nor lowercase) letters. \w
matches a word char.
However, best is to restrict it to [A-Z]+\.[A-Z]+
since you say the items you want to match follow the uppercase
+.
+uppercase
pattern.
See the regex demo
Upvotes: 2