by function then count_values creates wrong result.

[ghost]

I'm using the lahman2013.sqlite file to learn blaze. I believe the following is producing the wrong result.

import blaze as bz
_db = bz.Data('sqlite:///data/lahman2013.sqlite')

_sy=bz.by(_db.Teams.name, start_year=_db.Teams.yearID.min())
print(str(bz.compute(_sy, post_compute=False)))

produces

SELECT "Teams".name, min("Teams"."yearID") AS start_year 
FROM "Teams" GROUP BY "Teams".name

which seems correct to me.

I want to now use this result (the first year a team was in the MLB) to see how many teams entered the MLB for each year. I believe the following should produce that..

print(list(_sy.start_year.count_values())

but it produces incorrect counts. When I look at the sql produced I get the following:

print(str(bz.compute(_sy.start_year.count_values(), post_compute=False)))

SELECT anon_1.start_year, anon_1.count 
FROM (SELECT alias_1.start_year AS start_year, count(alias_2.start_year) AS count 
FROM (SELECT min("Teams"."yearID") AS start_year 
FROM "Teams" GROUP BY "Teams".name) AS alias_1, (SELECT alias_1.start_year AS start_year 
FROM (SELECT min("Teams"."yearID") AS start_year 
FROM "Teams" GROUP BY "Teams".name) AS alias_1) AS alias_2 GROUP BY alias_1.start_year) AS anon_1 ORDER BY anon_1.count DESC

which is incorrect. It should produce something close to:

select start_year, count(*)
from
 (SELECT "Teams".name, min("Teams"."yearID") AS start_year 
FROM "Teams" GROUP BY "Teams".name)
group by start_year

I would love to take a stab at fixing this myself. What I would like is for people to verify that I'm doing this correctly..

Thanks!

Billy

[cpcloud]

@Earney Thanks for the report! I have some thoughts on this, but I'm going to be out for the rest of the night. I'll post them first thing tomorrow.