Removing Multi-Column Statistics – A Process for Identification of Redundant Statistics

Database
The Teradata Database channel includes discussions around advanced Teradata features such as high-performance parallel database technology, the optimizer, mixed workload management solutions, and other related technologies.
Teradata Employee

Re: Removing Multi-Column Statistics – A Process for Identification of Redundant Statistics

Sivakumar,

That would makes sense to me if the additional non-PI column did not make any demographic changes, like all rows had the same status value and multi-column stats were on PI+Status.

I excluded the PI ones because I would not expect to see that situation often, for this particular reason.

Sometimes you will see PI+Partition, or PI+Partitioning columns. These do provide some specific information to the optimizer in some Teradata versions

Dave

Teradata Employee

Re: Removing Multi-Column Statistics – A Process for Identification of Redundant Statistics

Rahul,

I don not personally know the reason for this, but this is the way things work. For instance if you concatenate a few fields and one of them is NULL the result is also NULL.

This seems consistent with the multi-column stats rules

Dave

Enthusiast

Re: Removing Multi-Column Statistics – A Process for Identification of Redundant Statistics

Hi Dave, does this work on the Dieter's TD14 StatsInfo query?