Re: Are hash sub-partitions balanced?
From: <breitliw_at_centrexcc.com>
Date: Thu, 02 May 2013 09:29:54 -0600
Message-ID: <20130502092954.j8y6i96gowgwg488_at_webmail.telushosting.com>
Of course, if the data is very skewed then there is no way the distribution across the hash (sub)partitions can be even. Think of an extreme case. If there is just one distinct value for the partitioning column all rows will hash to the same partition and all the other partitions will receive 0 rows, no matter what hash function you apply.
Date: Thu, 02 May 2013 09:29:54 -0600
Message-ID: <20130502092954.j8y6i96gowgwg488_at_webmail.telushosting.com>
Of course, if the data is very skewed then there is no way the distribution across the hash (sub)partitions can be even. Think of an extreme case. If there is just one distinct value for the partitioning column all rows will hash to the same partition and all the other partitions will receive 0 rows, no matter what hash function you apply.
On Thu, 2 May 2013 11:23:54 -0400, Dba DBA
<oracledbaquestions_at_gmail.com> wrote:
Followup. So we loaded some skewed data and tihs is a rolliup from my
> sub-partitions spread across several partitions and I this does not appear
> to be very well balanced. Is this something I can expect? I am following
> the power of 2 with 16 hash sub-partitions. I am also going to get more
> data loaded.
> I just analyzed the table with granularity='ALL'
>
--
Wolfgang Breitling
Centrex Consulting Corporation
-- http://www.freelists.org/webpage/oracle-lReceived on Thu May 02 2013 - 17:29:54 CEST