Re: 'Range partitioning in PGSQL'

On Apr 11, 2011, at 4:04, Saseendra Babu wrote:

> ?'Range partitioning in PGSQL'
>
> Hello
> Is there as in oracle?
> My problem is as follows.
> I have telephone_ subscriber table with tel no as Pkey
> As the table has to contain millions of rows, I have to partition the table in to multiple partitions based on range of tel no.
> ( At present we did it with inheritance. but the problem found is: we have to insert rows in to each child table by referring the table name exclusively ,
> based on range of tel nos, keeping the parent table empty.)
> What we require is as follows.
> 1.. In the insert statement we shall specify only the table name ( not the partition name)
> 2.. The rows are be populated in to the corresponding partions.
> 3.. In select also we should only the table name ( not the partition name). rows are to be retrieved from corresponding partitions.
> 4.. While using partition by inheritance, we have to specify the partition, ie child table name.
> Waiting for your valued suggestions

You can write a trigger which transparently partitions inserts on the parent table into the appropriate partitions.

Michael Glaesemann
grzm seespotcode net