fix(sql): allow selecting aliased column followed by wildcard

[puzpuzpuz]

Fixes #2870

[bziobrowski]

Similar error still occurs if there's a join :

create table xo1 as 
(
  select
  rnd_double(0)*100 a,
  rnd_long() l,
  timestamp_sequence(0, 10000) k
  from long_sequence(3)
) timestamp(k);

create table xo2 as 
(
  select
  rnd_double(0)*100 a,
  rnd_long() l,
  timestamp_sequence(0, 10000) k
  from long_sequence(3)
) timestamp(k);


select cast(xo1.k as long), cast(xo2.k as long), *
from xo1  
join xo2 on xo1.k = xo2.k;

--or

select rnd_int()*k as zz, * 
from xo1  
join xo2 on xo1.k = xo2.k;

--Duplicate column [name=k]

[puzpuzpuz]

@bziobrowski good catch. Let me take a look at joins.

there is an issue to fix

[puzpuzpuz]

Similar error still occurs if there's a join

Addressed in 0087879

Note:

select rnd_int()*k as zz, *
from xo1  
join xo2 on xo1.k = xo2.k;**

This query is incorrect: should be rnd_int()*xo1.k or rnd_int()*xo2.k instead of rnd_int()*k.

[eugenels]

Similar error still occurs if there's a join :

create table xo1 as 
(
  select
  rnd_double(0)*100 a,
  rnd_long() l,
  timestamp_sequence(0, 10000) k
  from long_sequence(3)
) timestamp(k);

create table xo2 as 
(
  select
  rnd_double(0)*100 a,
  rnd_long() l,
  timestamp_sequence(0, 10000) k
  from long_sequence(3)
) timestamp(k);


select cast(xo1.k as long), cast(xo2.k as long), *
from xo1  
join xo2 on xo1.k = xo2.k;

--or

select rnd_int()*k as zz, * 
from xo1  
join xo2 on xo1.k = xo2.k;

--Duplicate column [name=k]

Should this work for cross join as well?

select *, xo1.k as kk1 from xo1 cross join xo2;
-- vs
select xo1.k as kk1, * from xo1 cross join xo2;

[puzpuzpuz]

Should this work for cross join as well?

Yes, it should. Fixed that in 3404ce6

[eugenels]

Looks good

[ideoma]

[PR Coverage check]

😍 pass : 8 / 8 (100.00%)

file detail

	path	covered line	new line	coverage
🔵	io/questdb/griffin/SqlOptimiser.java	8	8	100.00%