Re: Analytics updates - analyzing time to reload tables
- From: Nigel Thomas <nigel_cl_thomas@xxxxxxxxx>
- To: peter.schauss@xxxxxxx, oracle-l@xxxxxxxxxxxxx
- Date: Wed, 10 Oct 2007 14:03:50 -0700 (PDT)
Peter, are you really just dropping/creating indexes, or could you be
adding/removing constraints (eg unique/primary)? eg "alter table t1 add
constraint t_pk primary key (c1)"
creates an index to enforce the PK *and* modifies the last DDL time. You would
see that the table DDL time was the same as the last constraint/index to be
added.
Work through all other ALTER TABLE possibilities: are you using any of
ALTER TABLE T1 DISABLE LOCK
ALTER TABLE T1 DISABLE ALL TRIGGERS
or any other similar DDL on the table? These would all be reasonable things for
the load process to do (disable before load, re-enable afterwards...)
Regards Nigel
----- Original Message ----
From: "Schauss, Peter" <peter.schauss@xxxxxxx>
To: oracle-l@xxxxxxxxxxxxx
<smip>
I assume that the dba_objectslast_ddl_time for the table should be the
time that the table was truncated, having verified that truncating a
test table does, in fact, update this column. However, the
LAST_DDL_TIMEs for the tables are always later than the CREATED times
for the indexes. This would seem to say that the update process is
deleting and recreating the indexes before it updates the tables.
What am I missing here?
Thanks,
Peter Schauss
--
//www.freelists.org/webpage/oracle-l
Other related posts:
