There is no partitioning by size that I know of but at:
http://www.postgresql.org/docs/8.4/interactive/ddl-partitioning.html
there is very good documentation on the topic.
As of this last weekend I had myself to do some testing with
partitioning in Postgres 8.4. I had 7000 items. For each of them I
stored 6 variables (smallint) for each hour quarter. Test data was
generated for 5 years. That gave me 210336 records per id per 5 years
period and a total of 1,472,352,000 records. This is taking about 33MB
per partition table and a total of 231GB. Doing something like a
grouping by the id and summing up the values of 10 ids takes about 2.5
seconds, which looks to me like a quite a nice performance.
Petru Ghita
On 24/06/2010 15:05, Joshua Gooding wrote:
> rules for