XPath 2.0 regexp matching all strings except the words in a set

EastvanAxon · November 14th, 2012, 07:27 AM

Dear all,

I'm trying to write a regexp which allows me to find all words that are not in a set of words e.g.

if the inputs are "one" "two" "three" "four" "five" "bla" "blabla" I only want to obtain all of them except "two" and "three".

Obviously something like

Code:

[^two]|[^three]

does not work. I would need the 'and' operator, but it does not exist, right?

PS: I'm using the function "matches" to check the content of an attribute, like in:

Code:

[matches(@name,[^two]|[^three])]

Thanks in advance for your help!
EastvanAxon

mhkay · November 14th, 2012, 07:52 AM

I don't think regex is the right way to do this. If $stopwords is the set of words you want to exclude, and $input is the list of input words, then $input[not(. = $stopwords)] gives you the answer.

EastvanAxon · November 14th, 2012, 08:06 AM

Thanks for your answer.

I'm using regex because this is only a particular use case. Usually I set the value to just one name (because I want to get only strings matching that name).

here a piece of code which should be more explanatory:

Code:

<xsl:for-each select="[...]/function[matches(@name, document($config-file)/[...]/funcname)]">
	<xsl:variable name="currNode" select="position()"/>
	<xsl:variable name="uri" select= "[...]"/>
			
	<xsl:result-document href="{$uri}">
		<xsl:apply-templates select="/*">
			<xsl:with-param name="counter" select="$currNode"/>
		</xsl:apply-templates>
	</xsl:result-document>
</xsl:for-each>

where $configfile is a variable pointing to a file containing the regex. I believe I cannot use the solution you proposed because the file pointed by $configfile is meant to change, i.e. one can change the regex in order to change the result of the script, while the script must be general to address all cases and remain the same (selecting a particular function or all functions except some as in the example above).

I really hope this is possible. Otherwise I would like to know if there is another way to use the $configfile in order to be able to do the trick

Thanks again
EastvanAxon

mhkay · November 14th, 2012, 08:21 AM

I don't believe this can be done using XPath 2.0 regular expressions.

A more flexible solution would be for your external file to contain the rules in terms of XPath expressions (you would need to use an extension such as saxon:evaluate() to evaluate them).

EastvanAxon · November 14th, 2012, 09:00 AM

Thanks a lot. I'll try to think about the viability of employing external XPath expressions.

Best,
EastvanAxon