Hi Jim, With a boolean() function around 'tokenize' call, an "effective boolean value" of xs:string* result of 'tokenize' will be returned. But this would result in an error as per XPath 2.0 spec, for your example. I imagine, an 'exists' or 'count' function around tokenize() would let you achieve the same purpose as matches().
Sent from a mobile device. On Thursday, April 18, 2013, Jim Barnett <jim.barn...@genesyslab.com> wrote: > I’m having trouble getting tokenize() to behave the way I expect it to. As before, I have an element with an attribute @target with type idrefs. I want to split the value into the individual idrefs (I will run more complicated tests on them, but for the moment I’m just trying to get access to them). In the attached document and schema, I have two assertions. The first, using matches() succeeds, but the second, using tokenize(), fails: > > > > <xsd:assert test="matches(string(@target), '\S+(\s+\S+)+')" xpathDefaultNamespace="##targetNamespace"/> > > <xsd:assert test="boolean(tokenize(string(@target), '\s+'))" xpathDefaultNamespace="##targetNamespace"/> > > > > What am I doing wrong? > > > > Thanks, > > Jim -- Regards, Mukul Gandhi
