| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> Re: Consistent gets -- question
Hi Sat,
If ETL tool is doing bulk (or direct path) inserts (as it should be) and you can store 100 rows into the db block then you can have 0.01 block per row if no indexes/constraints on the table.
More indexes and constraints will give you more reads as in 1.
Regards,
Zoran Martic
"Oracle DBA"
> Hello All,
> I have a situation....
> We are running an etl process into a fact table has
> 15 fk's enabled (no
> indexes associated) During the session's run i did a
> trace session and
> this is the result from tkprof
>
> Execute 1175 23.09 23.22 10(disk)
> 3616(query)
> 1092396(current) 56400 (rows)
>
> As you can see there is a huge amount of current
> gets for 56400 rows.
>
> Now i removed all the constraints and ran the
> session again. this is the
> result..
>
> Execute 3520 29.60 29.02 0
> 6796(query)
> 14530 (current) 168960 (rows)
>
> The result shows very minimal current gets for a lot
> more rows than
> previous one.
>
> For some reason i am under the immpression that the
> current gets should
> at least match the number of rows inserted (or it is
> only for
> update/delete ??)
> The 2nd result disproves that..
> Please help me understand...
>
> Also i had seen in one of the metalink replies by
> the oracle tech that
>
> """the number of 'query' divided by the number of
> rows comes
> out to about 9 gets/row, which is not unreasonable.
> The number
> of 'current' divided by the number of rows is quite
> a bit higher
> at 33 gets/row, but may not be unreasonable,""""
>
> What is a resonable/unreasonale number for query
> get/row and current
> get/row
>
> Thanks,
>
> Sat
> --
> http://www.freelists.org/webpage/oracle-l
>
-- http://www.freelists.org/webpage/oracle-lReceived on Wed Feb 02 2005 - 09:09:20 CST
 |
 |