RE: Clustering factor smaller than table blocks.

  • From: "Christian Antognini" <Christian.Antognini@xxxxxxxxxxxx>
  • To: <Bernard.Polarski@xxxxxxxxxxxxxx>
  • Date: Wed, 27 Sep 2006 12:56:31 +0200


> If I read it correctly then a table whose number of blocks in PK
> inferior to number of blocks reported in dba_Tables had endure
> delete.

Two remarks:

1) "number of block" should be "number of referenced blocks".

2) As Jonathan already wrote, with ASSM there could be non-initialized
blocks below the HWM. In addition, with FSSM, since the HWM is increased
by 5 blocks at time (except for the first 5 blocks and when
_BUMP_HIGHWATER_MARK_COUNT has been set) you could have up to 5 empty
blocks ***per segment and free list group*** which are not referenced by
the index yet. 


Other related posts: