argmax aggregate function

Thanks Eugenia

Is there a way to create a feature request for this? I'm assuming UDTFS are not as fast as native functions. Am I right?

Best,
Amos.

Hi,

The thing is that this kind of aggregate function needs to materialize the second argument, only after we find the max of the first one, and not for every row. So I don't think is would be as efficient as possible.

In addition, from what I read in the docs, only one UDTF is allowed in the select clause, without anything else, which misses the whole point of issuing a a query such:

-- Getting the highest earning employee in each department
select argmax(salery, employee_name), department
from employees
group by department;

or even:
-- For each employee, get the salery diff to the department's senior
select employee_name, argmax(age, salery) over (partition by department) - salery as salery_diff_to_senior
from employees;

Is there a way keep track on this feature request?

Thanks,
Amos.