Hi Aron,

On Thu, May 19, 2005 at 12:41:15PM +0000, Aron Bock wrote:

> If <outcomes:columns> is the shorter list, you may want to iterate
> just that, at the appropriate junction.  Theoretically that should
> lessen wasted work.

I see---inside the template matching 'row', iterate over the values of
the 'attName' attributes in the 'column' elements and pull out the
values of the attributes with those names.  You're right, I suppose
that would be the minimum amount of work for a nested iteration like
this, but it's still O(n^2) in terms of pulling nodes out of the
source.

I guess I wanted something like this pseudocode:

if attribute-name is in list-of-attribute-names
  output attribute-value

Where list-of-attribute names is generated once in advance.  Then it
would be just O(n) where n is the number of rows.

> Alternatively you could let XSL do the lifting for you.  Something
> like so:
>
> With this imput (yours, simplified):
> 
> <data>
>    <outcomes>
>        <column attName="Computer" displayName="Computer Used" sum="no" />
>        <column attName="Location" displayName="The Location" sum="no" />
>    </outcomes>
>    <rows>
>        <row Computer="Rohan's PC" Location="IT Dept" Duration="555" />
>        <row Computer="Pharmacy" Location="Pharmacy" Duration="457" />
>    </rows>
> </data>
> 
> This XSL:
> 
> <?xml version="1.0" encoding="iso8859-1"?>
> <xsl:stylesheet version="1.0" 
> xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
>    <xsl:output method="text"/>
> 
>    <xsl:template match="/">
>        <xsl:for-each select="data/rows/row">
>            <xsl:for-each select="@*[name() = 
> /data/outcomes/column/@attName]">
>                <xsl:value-of select="."/>
>                <xsl:text>, </xsl:text>
>            </xsl:for-each>
>            <xsl:text>&#xa;</xsl:text>
>        </xsl:for-each>
> 
>    </xsl:template>
> 
> </xsl:stylesheet>
> 
> Produces:
> 
> Rohan's PC, IT Dept,
> Pharmacy, Pharmacy,

That's getting closer to what I want, at least in terms of neatness.
I'm just wondering if we can get rid of the nested for-each.

> I suppose you could use keys for faster access, but I don't know how
> that would affect total transform time.

You're right, I guess the question is "does this affect the transform
time?"  Even the nested for-each constructs don't seem to, probably
because the input is only of the order of hundreds or a few thousand
rows in my case.  I was just wondering whether there is an idiom in
XSL that is analogous to my pseudocode above.  I can't seem to hit on
it myself.


-- 
Paul.

w  http://logicsquad.net/
h  http://paul.hoadley.name/