Re: [xsl] things about grouping

[Ihe Onwuka <ihe.onwuka@xxxxxxxxx>]

On Thu, Nov 29, 2012 at 8:25 AM, Ihe Onwuka <ihe.onwuka@xxxxxxxxx> wrote:
> On Thu, Nov 29, 2012 at 8:22 AM, Ihe Onwuka <ihe.onwuka@xxxxxxxxx> wrote:
>> On Thu, Nov 29, 2012 at 8:18 AM, Ihe Onwuka <ihe.onwuka@xxxxxxxxx> wrote:
>>> On Thu, Nov 29, 2012 at 8:15 AM, Andrew Welch <andrew.j.welch@xxxxxxxxx> wrote:
>>>>> the comma operator corresponds to union written as | (by definition a
>>>>> set is unordered so the additional ordering constraint imposed by the
>>>>> , operator is meaningless)
>>>>
>>>> The comma is just a way of separating items in sequence, whereas union
>>>> de-dupes and sorts into document order.
>>>>
>>>> select="@foo, @foo, @foo"
>>>>
>>>> vs
>>>>
>>>> select="@foo | @foo | @foo"
>>>>
>>>> the former selects a sequence of 3 items, the latter a sequence of 1.
>>>>
>>>
>>> ok but I'm translating , into a set-theoretic construct, in which
>>> domain a,b is indistinct from a|b - correct me if I have got that
>>> wrong.
>>
>> sorry you mean the deduping........
>>
>> well you are no longer dealing in sets ...... so if you apply a
>> set-theoretic construct then all bets are off and the law of the
>> jungle applies.
>
> or to put it another way that is as much type abuse as applying a
> numeric operator to a string and for that reason I would walk away
> from such expressions.

unless the , operator results in a relation rather than a set but then
I'd have to refer to my discrete math text and parse the effect of a
set-theoretic difference operation on a relation if indeed that is
valid thing to do.