Re: [xsl] xsl:analyze-string

Subject: Re: [xsl] xsl:analyze-string
From: "M. David Peterson" <m.david@xxxxxxxxxx>
Date: Mon, 02 Aug 2004 02:02:23 -0600
Hey David,

While not a complete answer to your problem I wonder if this might be of help to you...

http://blogs.msdn.com/the1/archive/2004/03/29/102123.aspx

This is a coding challenge called "Phraser" put up by Zhanyong Wan whom if you are not already familiar with comes from the Haskell school of thought (his dissertation for his Ph.D. in Computer Science from Yale was "Functional Reactive Programming for Real-Time Reactive Systems".

The Challenge:

In his words... "Your task is to write a program that finds all possible ways to phrase a phone number. By "phrase" I mean using a mixture of numbers and English words to spell out the number. For example, the number 642-394-6369 can be phrased as nice-window, nice-wind-ox, nice-win-fox, nice-9-in-fox, etc."

He created his recursive solution using Haskell and I wonder if his functional-programming-based algorithm (he writes a fantastic explanation regarding the thought process behind his solution) might help you solve the problem you are running into.

I've been meaning to attempt an XSLT solution to his challenge and if I can ever find the time I will most definitely make an attempt... None-the-less, thats of know help to you now! :) But maybe this is close enough to your issue that it may be of some use...

HTH's!

Best regards,

<M:D/>


David.Pawson@xxxxxxxxxxx wrote:
Problem:

I have xml looking like
   <record>19/08/1992  09:08            111938 500014.08b</record>

Of interest is the last field, the dos file name.
Its general form is
nnnnnn.nnA   n=int, A =A-Z.

I can use the analyze-string to parse this (mostly),
I'm missing a processing model to take care of exceptions.

E.g. xxx.01a is a particular (valid form).
On occasions someone has misnamed the file xxx.ola  using o instead
of 0 letter l instead of 1.

I'm missing the nest of cases: My code is shown below.
What's the logic to posively select all the known/valid combinations
are exhausted?
I want to add some form of error markup to identify these rogues.

Valid options include, for regex-group(5)

nnb
nnm
nnl
nnf
lab

TIA, regards DaveP



<xsl:template match="record">
  <xsl:analyze-string  select="." flags="i" regex=
"([0-9]+/[0-9]+/[0-9]+) +([0-9]+:[0-9]+) +([0-9]+) +([0-9]+)\.([0-9a-z]+)" >
    <xsl:matching-substring>
      <file>
        <nm><xsl:value-of select="regex-group(4)"/></nm>
        <ext><xsl:value-of select="regex-group(5)"/></ext>
        <xsl:analyze-string select="regex-group(5)" flags="i" regex=
"([0-9]{{1,2}})([a-z])">
          <xsl:matching-substring>
            <vol><xsl:value-of select="regex-group(1)"/></vol>
            <type>
              <xsl:choose>
                <xsl:when test="regex-group(2) = 'b' or regex-group(2) =
'B'">
                  <xsl:text>braille</xsl:text>
                </xsl:when>
                <xsl:when test="regex-group(2) = 'm' or regex-group(2) =
'M'">
                  <xsl:text>Moon</xsl:text>
                </xsl:when>
                <xsl:when test="regex-group(2) = 'l' or regex-group(2) =
'L'">
                  <xsl:text>ATOD</xsl:text>
                </xsl:when>
               <xsl:when test="regex-group(2) = 'f' or regex-group(2) =
'F'">
                  <xsl:text>DiskFile</xsl:text>
                </xsl:when>
                <xsl:otherwise>
                  <xsl:value-of select="."/>
                </xsl:otherwise>
              </xsl:choose>
            </type>
          </xsl:matching-substring>
        </xsl:analyze-string>

          <xsl:if test="regex-group(5)='lab'">
            <type>label</type>
          </xsl:if>
    </file>
    </xsl:matching-substring>
  </xsl:analyze-string>
</xsl:template>


Regards DaveP.


**** snip here *****

Current Thread