[Mondrian] Unexpected behaviour while using a TopCount in a Named Set

Julian Hyde jhyde at pentaho.com
Wed Jul 11 15:06:04 EDT 2012


On Jul 11, 2012, at 11:28 AM, Pedro Alves wrote:

> does this also apply for native top count evaluation?

No.

In May, Pedro Vale fixed native evaluation of topcount so that it took ranges in the slicer into account, but he didn't fix non-native evaluation.

This bug is that non-native evaluation now behaves differently to native evaluation. Well, duh.

Adding Distinct( ... ) has the effect (the only effect, in this case) of disabling native evaluation.

A workaround, obviously, is to make the argument to TopCount simple enough that it can use native evaluation.

Julian



More information about the Mondrian mailing list