Hello!
I'm new to XSL/Xpath and would be glad if you could give me any advice!

A client of ours uses a program that generates a text file, whose fields are added in an XML structure. One of the nodes have the following information:

<MsgCantoInfEsq>@ Milhas }]# (Inlcuindo o limite de Empréstimo Fácil) ~</MsgCantoInfEsq>

I need to make this replacement:
- from '@' to <h1 class="label">
- from '}' to </h1>
- from ']' to <p class="espaco">
- from '#' to <p>
- from '~' to </p>


First, I tried using 'replace()'. For instance, replace(//MsgCantoInfEsq, '#', '&lt;p>'). Still, when the browser renders the code, it does not apply the style for this tag.

Then, I tried the following:

<xsl:variable name="msg" select="tokenize(//MsgCantoInfEsq, '\s+')"/>
<xsl:for-each select="$msg">
    <xsl:choose>
        <xsl:when test="$msg = '@'">
            <xsl:text><![CDATA[<h1 class="label">]]></xsl:text>
        </xsl:when>
        <xsl:when test="$teste = '}'">
            <xsl:text><![CDATA[</h1>]]></xsl:text>
        </xsl:when>
        <xsl:when test="$teste = ']'">
            <xsl:text><![CDATA[<p class="espaco">]]></xsl:text>
        </xsl:when>
        <xsl:when test="$teste = '#'">
            <xsl:text><![CDATA[<p>]]></xsl:text>
        </xsl:when>
        <xsl:when test="$teste = '~'">
            <xsl:text><![CDATA[</p>]]></xsl:text>
        </xsl:when>
        <xsl:otherwise>
            <xsl:value-of select="."/>
        </xsl:otherwise>
    </xsl:choose>
</xsl:for-each>

I was expecting to get this:
<h class="label">Milhas</h>
<p class="espaco">
<p>(Inlcuindo o limite de Empréstimo Fácil)</p>

However, nothing is displayed. I'm using Saxon 8.

Could you please help me??

Thanks in advance!

Whoever designed that format shouldn't be allowed near XML.
One of the problems is that you are creating text, not markup, so if you look at the raw output you will see the angle brackets are escaped.
In my opinion you'd be better off pre-processing with a text tool to replace the special symbols with there true counterparts but someone else may have a clever way of doing what you need from within XSLT.

--

Joe (Microsoft MVP - XML)

XSLT constructs a result tree which can contain element nodes e.g. h1 or p element nodes and these element nodes can have attribute nodes (like class). It is not working on a text level where you create start tags like <h1> or end tags like </h1>.
So you need to change your approach and create elements like h1 or p either as literal result elements
or with xsl:element
unless you want to resort to disable-output-escaping hacks.


--
  Martin Honnen
  Microsoft MVP - XML

Your mistake is in the title of your post. You are not trying to replace the characters with strings, you are trying to replace them with element nodes. XSLT generates a tree of nodes, not lexical XML - never forget this!

Since you're using XSLT 2.0, the answer lies with xsl:analyze-string, which allows new element nodes to be constructed within the <xsl:matching-substring> instruction.

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

All the above advice is correct, but on a completely different level, your RegEx is incorrect, as is your for-each loop.

$msg will be an array containing the following values:

None of the values in the array equals ']' or '}' or '#'.

Also, in your for-each loop the first test is testing $msg - which is the whole array. You then proceed to test $teste - a variable I see no declaration for (why this doesn't give you an error I don't know).



/- Sam Judson : Wrox Technical Editor -/

I really appreciate all the replies!! Since Friday, I've been analyzing each one of them in order to better understand our options here. Also, I've been checking this site http://www.dpawson.co.uk and getting some ideas!

Sam, this may sound stupid, but could you please tell me why in my for-each loop the first test is testing the whole array, and what changes I have to do in order to start evaluating the first position instead of the entire array??
I tried another test, just to better understand how for-each works:

<xsl:variable name="teste" select="tokenize('tchau ola tchau', '\s')"/>
<xsl:for-each select="$teste">
    <xsl:choose>
        <xsl:when test="$teste = 'ola'">
            <xsl:text>Achei ola!</xsl:text>
        </xsl:when>
        <xsl:when test="$teste = 'tchau'">
            <xsl:text>Achei tchau!</xsl:text>
        </xsl:when>
        <xsl:otherwise>
            <xsl:text>Nao nada!</xsl:text>
        </xsl:otherwise>
    </xsl:choose>
</xsl:for-each>

I thought the output would be something like:
Achei tchau!
Achei ola!
Achei tchau!

However, I get:
Achei ola!
Achei ola!
Achei ola!

Btw, you were correct with the $test variable; when I posted the first message, I forgot changing '$teste' to '$msg'. Sorry!

I think you want
--
  Martin Honnen
  Microsoft MVP - XML

If $teste is an array, then $teste = 'X' test to see if ANY element in the array equals 'X', i.e. in your example above for every loop through the for-each you are testing the WHOLE array again.

You want this:

/- Sam Judson : Wrox Technical Editor -/

Thanks guys, I *really* appreciate your help!! And I told you the question was stupid..... sorry!!!