RE: using xsl:for-each-group, grouping by child nodes?

Of course it's possible, and not really difficult. 

I have to say my instinct would be to start by transforming this into more
XML-like XML, something like

<players>
  <player>
    <name>joe</name>
    <age>2</age>
    <position>fielder</position>
    <state>ny</state>
    <team>mets</team>

because all subsequent processing you do on the data would then be a lot
clearer. (I know the number isn't the player's age, but I have no idea what
it actually is...)

With that it would be

<xsl:template match="players">
  <xsl:for-each-group select="player" group-by="concat(team, state)">
     <xsl:value-of select="current-grouping-key()"/>
     <xsl:for-each-group select="current-group()" group-by="position">
       <...

Now, if you don't want to do the preprocessing, you just substitute
different expressions into the above:

players -> RESULTSET
player -> ROW
team -> COL[5]/DATA
state -> COL[4]/DATA

etc.

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

> -----Original Message-----
> From: Eric Scheid [mailto:eric.scheid@xxxxxxxxxxxxxxx] 
> Sent: 04 October 2008 08:39
> To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> Subject:  using xsl:for-each-group, grouping by child nodes?
> 
> 
> I have export data that looks like this:
> 
> <RESULTSET FOUND="6">
>     <ROW MODID="3" RECORDID="1">
>         <COL><DATA>joe</DATA></COL>
>         <COL><DATA>2</DATA></COL>
>         <COL><DATA>pitcher</DATA></COL>
>         <COL><DATA>ny</DATA></COL>
>         <COL><DATA>mets</DATA></COL>
>     </ROW>
>     <ROW MODID="3" RECORDID="2">
>         <COL><DATA>mark</DATA></COL>
>         <COL><DATA>11</DATA></COL>
>         <COL><DATA>outfielder</DATA></COL>
>         <COL><DATA>ny</DATA></COL>
>         <COL><DATA>mets</DATA></COL>
>     </ROW>
>     <ROW MODID="3" RECORDID="3">
>         <COL><DATA>jane</DATA></COL>
>         <COL><DATA>13</DATA></COL>
>         <COL><DATA>outfielder</DATA></COL>
>         <COL><DATA>ny</DATA></COL>
>         <COL><DATA>cubs</DATA></COL>
>     </ROW>
>     <ROW MODID="3" RECORDID="4">
>         <COL><DATA>mike</DATA></COL>
>         <COL><DATA>7</DATA></COL>
>         <COL><DATA>outfielder</DATA></COL>
>         <COL><DATA>ny</DATA></COL>
>         <COL><DATA>mets</DATA></COL>
>     </ROW>
> </RESULTSET>
> 
> I'm trying to transform it into this:
> 
>     Team: mets, ny
>         pitchers:
>             joe, 2
>         oufielders:
>             mark, 11
>             mike, 7
> 
>     Team: cubs, ny
>         outfielders:
>             jane, 13
> 
> .. so naturally I thought of using for-each-group, only 
> problem is I can't find any examples which shows how to 
> group-by using values in the child nodes .. so I don't know 
> if this is even possible, let alone exactly how to specify 
> it. I've tried a few different things to no avail.
> 
> Is this even possible?
> 
> e.
> 
> 
> --
> Ironclad Networks
> Information Architecture
> http://www.ironclad.com.au/
> 
> "Providing tactical IA services to web agencies"