Bug #80116 Outer reference to aggregate incorrectly 0 or NULL in subquery
Submitted: 22 Jan 2016 13:56
Reporter: Knut Anders Hatlen Email Updates:
Status: Verified Impact on me:
None 
Category:MySQL Server: Optimizer Severity:S3 (Non-critical)
Version:5.7 OS:Any
Assigned to: CPU Architecture:Any

[22 Jan 2016 13:56] Knut Anders Hatlen
Description:
If a subquery in the select list or the order by clause has a reference to an aggregate in the outer query, that reference seems to evaluate to 0 or NULL instead of the correct value.

How to repeat:
create table t(a int);
insert into t values (1), (2), (3);

# returns (6, NULL) instead of the expected (6, 1)
select sum(a) as s, (select 1 having s) from t;

# returns (6, 1) instead of the expected (6, NULL)
select sum(a) as s, (select 1 having s is null) from t;

# returns (3, NULL) instead of the expected (3, 1)
select count(a) as c, (select 1 having c) from t;

# returns (3, 1) instead of the expected (3, NULL)
select count(a) as c, (select 1 having c = 0) from t;

# returns the results in the wrong order.
# expected result: (3, 1), (2, 1), (1, 1)
# actual result: varies, have seen both
#       (1, 1), (2, 1), (3, 1)
# and   (3, 1), (1, 1), (2, 1)
select a, count(*) as c from t group by a order by (select -a having c > 0);