cancel
Showing results for 
Search instead for 
Did you mean: 
srini_ch
Astronaut
Status: New

Suppose we join a child table and parent table in Incorta based on a column that is not unique for both child and parent tables. Incorta join by default will pull only one row from parent table even though it has multiple rows satisfying that join condition. But the same functionality is exhibited if there is a join between the tables in Incorta and when the SQL query is ran from dbeaver (3rd party software) with same conditions using left join. And the same SQL query returns multiple rows from parent table when the join in Incorta is removed. Some where the join functionality is enforced when Incorta rewrites the query.

1 Comment
RADSr
Captain
Captain

I think you can do that if you go to the spark port rather than the SQLi/SQLx port?   

It would make sense that the Incorta engine would enforce Incorta joins when it translates the query, by bypassing it you'd lose query performance ( speed )  but gain more control since you're basically telling spark to join specific parquet files.

Full disclosure:  I could be completely off base  😉    but I think it's worth exploring both external tool connection string options.