I’m currently trying to improve access to our hadoop connections. Our impala table is partitioned, in product and month. So the idea is, if I know which product I’d like to investigate and maybe also the date it was processed, I could make a preselection on those partitions. I understand it is fairly fast to show partitions in impala by running this query:
SHOW PARTITIONS my_table
Unfortunately this is not possible in Spotfire, since the show statement is not supported. Sure, I can obtain the same result running this query:
SELECT DISTINCT product, month FROM my_table
However, this is rather slow and I expect Spotfire to be able to expose partition keys (columns) automatically as prompts to end users in the most performant way.