Hi Martin, happy new year to you too :-)
"My question is, does it make sense to manually create extended statstics
for two columns, that are already part of a two-column index?"
It matters to create extended statistics if there is a big estimation
error even you have multicolumn index. In the case of absence of extended
statistics the optimizer tries to use DISTINCT_KEYS of the multi column
index in order to estimate cardinality. So, if your data (column) is highly
skewed in this case the multicolumn index may not help optimizer well
enough to calculate cardinality properly, Therefore , you can create
extended statistics to solve the problem. But you can keep multi column
index in order to retrieve data efficiently.
Thanks
On Wed, Jan 9, 2019 at 2:26 PM Martin Klier - Performing Databases GmbH <
martin.klier@xxxxxxxxxxxxxxxxx> wrote:
Hi listers,
happy new year to all of you!
I was wondering if it is useful to still have manually created extended
statistics (to enhance cardinality calculation for logically connected
columns) and multi-column indexes of the same columns at the same time.
I need the two- or three-column indexes for query performance reasons
anyway, so leaving them out is not an option.
My question is, does it make sense to manually create extended statstics
for two columns, that are already part of a two-column index?
My field of using this stuff is versions 11.2, 12.1, 12.2 and 18.3 on
Linux.
Background: I can't see an extension in dba_stat_extensions if I do not
have the extended stats created, so I was thinking, if the CBO can
calculate the correct cardinality from the index / index stats, I could
spare my stats jobs the effort to calculate many of my virtual column stats.
Thank you in advance, and please don't laugh if I am way off the path... :)
Best regards
--
Martin Klier // Performing Databases GmbH
Managing Partner // Senior DB Consultant
Oracle ACE Director
martin.klier@xxxxxxxxxxxxxxxxx // https://www.performing-databases.com
