Obtaining the max value out of a series.

BillMoo · June 29th, 2012, 04:16 AM

Hello,

I am trying to create a graph of my data using <svg> but in order to scale the y axis correctly I need to find the max of a series of numbers.

To do this I thought the best approach would be to use something like this:

Code:

<!-- Plot Bars -->
<xsl:variable name="maxY" as="xs:decimal" select="0"/>
<xsl:for ...>
    ...
    <xsl:if test="@attr_1 >= $maxY">
        <xsl:value-of select="$maxY = @attr_1"/>
    </xsl:if>
    ...
</xsl:for>
<!-- Use maxY here to plot Y axis -->

But this does nothing, I'm left with 0, so I then thought:

Code:

<!-- Plot Bars -->
<xsl:variable name="maxY" as="xs:decimal" select="0"/>
<xsl:for ...>
    ...
    <xsl:value-of select="$maxY = max($maxY, @attr_1)"/>
    ...
</xsl:for>
<!-- Use maxY here to plot Y axis -->

But this produces and error during compilation of : Required item type of second argument of max() is xs:string; supplied value has item type xs:decimal

My XSL is Schema aware so it knows that @attr_1 really is a xs:decimal so needless to say I am quite confused with this as the W3C docs imply that it will return the max of a series of comma limited entries!

So I would appreciate your thoughts on this please.

--
Bill

samjudson · June 29th, 2012, 04:33 AM

You can't update a variable in XSLT - its a functional programming language.

The following should give you the max of all attr_1 attribute values of the child element.

<xsl:variable name="maxY" select="max(element/@attr_1)"/>

BillMoo · June 29th, 2012, 04:45 AM

Quote:

Originally Posted by samjudson

You can't update a variable in XSLT - its a functional programming language.

The following should give you the max of all attr_1 attribute values of the child element.

<xsl:variable name="maxY" select="max(element/@attr_1)"/>

And indeed it does. Thank you very much, I'm not making excuses for myself but as a C++ programmer my approach was logical and sane.

--
Bill

mhkay · June 29th, 2012, 05:43 AM

The actual error message is because the first argument of max() must be a sequence, and the second (optional) argument is a collation to be used for the comparison. So to get the max of X and Y, you need max((X,Y)) rather than max(X,Y) - if you use the latter form, Y is taken as the collation.

But as Sam has explained, the underlying problem in your code is deeper: you want to compute the max by supplying the whole sequence to the function, not by computing the max of each pair of values in turn.