Re: Mention ordered datums in PartitionBoundInfoData comment

On Fri, Dec 8, 2017 at 12:25 AM, Robert Haas <robertmhaas@gmail.com> wrote:
> On Tue, Dec 5, 2017 at 1:03 AM, Ashutosh Bapat
> <ashutosh.bapat@enterprisedb.com> wrote:
>> Sorry. Thanks for pointing it out. fixed in the attached patch.
>
> + * The datums in datums array are arranged in the increasing order defined by
>
> Suggest: in increasing order as defined
>

Done.

> There's a second place where the same change is needed.

Done.

>
> + * resp. For range and list partitions this simply means that the datums in the
>
> I think you should spell out "respectively" instead of abbreviating to "resp".

Done.

>
> + * datums array are arranged in the increasing order defined by the partition
> + * key collation.
>
> It's not just the collation but also, and I think more importantly,
> the operator class.   And there can be multiple columns, and thus
> multiple opclases/collations.  Maybe "defined by the partition key's
> operator classes and collations".

I had forgot about the operator class. Sorry. Done.

>
> + * PartitionBoundInfoData structures for two partitioned tables with exactly
> + * same bounds look exactly same.
>
> This doesn't seem to me to add much.
>

We have a comment in partition_bounds_equal()'s prologue

"PartitionBoundInfo is a canonical representation of partition bounds.".

But we may use that property in other places also, so having it in
prologue of PartitionBoundsInfoData makes sense. For now, I have
removed those lines.

PFA updated patch.
-- 
Best Wishes,
Ashutosh Bapat
EnterpriseDB Corporation
The Postgres Database Company