hello @Jackie, when i change my query from a normal

userId

to

ID_SET(userId)

, it times out for me on the quickstart docker image

# without id_set
SELECT userId, count(userId) FROM events WHERE ((timeBin BETWEEN 0 AND 21) AND (cell BETWEEN 1000 AND 1005)) GROUP BY userId having count(*) > 15

# with id_set
Timed out while combining group-by order-by results after 9995ms

my field

userId

are comprised of extremely long strings (e.g.

F0B19EF3B76702507C5C07592148556B9808D3E5E49344FD4B162D51BEFF60B3

), would it help to change them to ints? alternatively, i’ve tried the following

# get the distinct set of user ids
select ... from events where ...

# manually form the sql query from the result of previous query
select ... from events where userIds in $RESULT_OF_FIRST_QUERY and $OTHER_CONDITIONS

this works, with up to 1000

userId

. but it fails with

ID_SET

another use case - how do i get the

ID_SET

when i want to filter with a group by + having clause? e.g.

SELECT user, count(*) FROM events WHERE time BETWEEN 0 AND 31 AND location BETWEEN 1000 AND 1005 GROUP BY user HAVING count(*) > 10

user	count(*)
1052157	15
97517	17
...

SELECT ID_SET(user), count(*) FROM events WHERE time BETWEEN 0 AND 31 AND location BETWEEN 1000 AND 1005 GROUP BY user HAVING count(*) > 10

idset(user)	count(*)
AgAAAAABAAAAADowAAABAAAAEAAAABAAAAD9DQ==	15
AgAAAAABAAAAADowAAABAAAAAQAAABAAAADtfA==	17
...

for future ref, seems like i would have to create the

IdSet

on the client side, so client would have to manage 2 different queries