Re: Speeding up INSERTs and UPDATEs to partitioned tables

On 1 November 2018 at 06:45, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> On Wed, Aug 22, 2018 at 8:30 AM David Rowley
> <david(dot)rowley(at)2ndquadrant(dot)com> wrote:
>> On 22 August 2018 at 19:08, Amit Langote <Langote_Amit_f8(at)lab(dot)ntt(dot)co(dot)jp> wrote:
>> > +#define PartitionTupRoutingGetToParentMap(p, i) \
>> > +#define PartitionTupRoutingGetToChildMap(p, i) \
>> >
>> > If the "Get" could be replaced by "Child" and "Parent", respectively,
>> > they'd sound more meaningful, imho.
>>
>> I did that to save 3 chars. I think putting the additional
>> Child/Parent in the name is not really required. It's not as if we're
>> going to have a ParentToParent or a ChildToChild map, so I thought it
>> might be okay to assume that if it's "ToParent", that it's being
>> converted from the child and "ToChild" seems safe to assume it's being
>> converted from the parent. I can change it though if you feel very
>> strongly that what I've got is no good.
>
> I'm not sure exactly what is best here, but it seems to unlikely to me
> that somebody is going to read that macro name and think, oh, that
> means "get the to-parent map". They are more like be confused by the
> juxtaposition of "get" and "to".
>
> I think a better way to shorten the name would be to truncate the
> PartitionTupRouting() prefix in some way, e.g. dropping TupRouting.

Thanks for chipping in on this.

I agree. I don't think "TupRouting" really needs to be in the name.
Probably "To" can also just become "2" and we can put back the
Parent/Child before that.

I've attached v11, which does this.

--
David Rowley http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services