RE: problem with tokenize

Try:

<xsl:analyze-string regex='"[^"]*"|[^,]*'>
  <xsl:matching-substring>
    <token>
      <xsl:value-of select='translate(, """", "")'/>
    </token>

Not tested.

Michael Kay
http://www.saxonica.com/

> -----Original Message-----
> From: Kent Seegmiller [mailto:hookjaw20@xxxxxxxxxxx] 
> Sent: 10 February 2006 07:16
> To: XSLT
> Subject:  problem with tokenize
> 
> I am trying to tokenize a text file using 2.0:
> 
> 10/12/05,"Caldwell, ID",205,"12,569",625,8
> 10/23/05,"Seattle, WA","1,092","1,269",695,4
> 10/31/05,"Denver, CO",203,"3,567","9,668",7
> 
> but I don't want to split words or numbers inside quotation marks.
> Would it be easier to use recursion, or tokenize?
> 
> Thanks in advance