Re: [POC] hash partitioning

On Tue, Jun 6, 2017 at 1:03 PM, amul sul <sulamul(at)gmail(dot)com> wrote:
> May I ask you, how you sure about 8 is an unfit value for t1 relation?
> And what if the value other than 8, for e.g. 7?

Well, First I created t1 as a leaf relation like below, and I tested
insert into t1 with value 8 and it was violating the partition
constraint of t1, however, 7 was fine.

create table t (a int) partition by hash(a);
create table t1 partition of t for values with (modulus 2, remainder 1);

Later I dropped this t1 and created 2 level partition with the leaf as a range.

drop table t1;
create table t1 partition of t for values with (modulus 2, remainder
1) partition by range(a);
create table t1_1 partition of t1 for values from (8) to (10);

So now, I am sure that t1_1 can accept the value 8 and its parent t1 can't.

So I think this can only happen in the case of partitioned by hash
that a value is legal for the child but illegal for the parent? Isn't
it a good idea that if a user is inserting in the top level relation
he should know for which partition exactly the constraint got
violated?

>
> Updated patch attached.

Thanks.

--
Regards,
Dilip Kumar
EnterpriseDB: http://www.enterprisedb.com