Random ordering

This is why there is a need for random ordering:
If you want a list of items and for any reason you want to make sure that the ordering doesn't follow any logic pattern. For example, you 
want to show a set of products but don't want to reveal how their prices differ, or when they were introduced.
Simply getting it in document order is not good enough, because the nodes was probably inserted at the first time with some sort of order. 
But more important, they would always appear in the same order each time you repeat the selection.

The second case extends the first one, you would want one or more items at random.
So you select a nodest with random ordering, then only returns the first (n) items.
It could be used for example like this: "click here to get to a random product of our vast selection". Or "click here to find your blind 
date".
As in the first example, the resultset shouldn't repeat if you do another selection.
Everytime you select, there should be a new product returned.

So I understand RANDOM is not in the SQL standard. But even so there is obviously a need for it, so many RDBMS have implemented it anyway. 
That is an indication that it may be needed in XQuery too.

Received on Thursday, 31 May 2007 13:52:37 UTC