not handling usage of group function and window function at some conditions

Hi,

looks like SparkSQL is not able to handle this query:

SELECT rank() OVER (ORDER BY 1), count(*) FROM empsalary GROUP BY 1;

PgSQL, on the other hand, does.