You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Similar to #3873, it is debatable whether the minimum of 1.0 and 1.00 should be the former or latter. Do we also need to output the canonical value from the equivalent class so that it is agnostic of input order?
To Reproduce
create table t (v1 decimal, v2 interval);
create materialized view mv1 as select min(v1) as min1, min(v2) as min2, max(v1) as max1, max(v2) as max2 from t;
insert into t values (1.0, '30 day'), (1.00, '1 month');
select * from mv1;
select min(v1) as min1, min(v2) as min2, max(v1) as max1, max(v2) as max2 from t;
delete from t where v1 = 1;
insert into t values (1.00, '1 month'), (1.0, '30 day');
select * from mv1;
select min(v1) as min1, min(v2) as min2, max(v1) as max1, max(v2) as max2 from t;
Expected behavior
Open to discuss. Possible behaviors:
First value win / last value win (current and PostgreSQL)
Canonical value from the equivalent class
order by
Additional context
No response
The text was updated successfully, but these errors were encountered:
This is just one of the indeterminism sources. No immediate fix needed. Keeping it open for more time so that we can summarize our decision regarding such cases.
Describe the bug
Similar to #3873, it is debatable whether the minimum of
1.0
and1.00
should be the former or latter. Do we also need to output the canonical value from the equivalent class so that it is agnostic of input order?To Reproduce
Expected behavior
Open to discuss. Possible behaviors:
Additional context
No response
The text was updated successfully, but these errors were encountered: