On Fri, May 8, 2020 at 7:27 AM Andy Fan <zhihui.fan1213@gmail.com> wrote:
>> + else if (inner_is_onerow)
>> + {
>> + /* Since rows in innerrel can't be duplicated AND if
>> innerrel is onerow,
>> + * the join result will be onerow also as well. Note:
>> onerow implies
>> + * multi_nullvals = false.
>>
>> I don't understand this comment. Why is there only one row in the other
>> relation which can join to this row?
>
>
> I guess you may miss the onerow special case if I understand correctly.
> inner_is_onerow means something like "SELECT xx FROM t1 where uk = 1".
> innerrel can't be duplicated means: t1.y = t2.pk; so the finally result is onerow
> as well. One of the overall query is SELECT .. FROM t1, t2 where t2.y = t2.pk;
>
>
> I explained more about onerow in the v7 README.unqiuekey document, just copy
> it here.
For some reason this mail remained in my drafts without being sent.
Sending it now. Sorry.
My impression about the one row stuff, is that there is too much
special casing around it. We should somehow structure the UniqueKey
data so that one row unique keys come naturally rather than special
cased. E.g every column in such a case is unique in the result so
create as many UniqueKeys are the number of columns or create one
unique key with no column as you have done but handle it more
gracefully rather than spreading it all over the place.
Also, the amount of code that these patches changes seems to be much
larger than the feature's worth arguably. But it indicates that we are
modifying/adding more code than necessary. Some of that code can be
merged into existing code which does similar things as I have pointed
out in my previous comment.
Thanks for working on the expanded scope of the initial feature you
proposed. But it makes the feature more useful, I think.
--
Best Wishes,
Ashutosh Bapat