[postgis-users] Partitions

[Andy Colson]

On 3/1/2013 12:25 PM, oisink wrote:
> Hi
>
> I'm about to build a large spatial database (OS MasterMap - ~440M features).
> I intend constraining the partitions against 25km grid squares. However,
> using this approach some features will inevitably cross the boundary between
> my partition constraints. Can this be done?
>
> How does it work in practice, i.e. when loading data do I need to check I am
> not putting 'trans-partition' features into more than one partition, or do I
> duplicate them across partitions?
>
> Regards
>
> O
>

Are you going to be deleting and re-adding huge chunks of data?  I 
believe that's what partitions are best for.  If your dataset is mostly 
read-only, then partitions wont help.

 From what I've read, you should not duplicate, otherwise if you "select 
* from parent" you will get duplicate records back.

-Andy