[xquery-talk] Count of Distinct elements performance problem

[Michael Kay]

> 	Our requirement is we need to parse the input XML and 
> get the Count, NullCount, MaxLength, DistinctCount etc and 
> apply some business rules on the results. That is it. We do 
> not want to store this data in Database as we don't need this 
> XML in future.

Frankly, for doing a fairly simple aggregation like this of a very large
dataset, in which it's not possible to amortize the cost of indexing over
many queries, I think I would write a SAX application in Java.

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