What is the best practice on to dictionary or not to dictionary string id columns with billions of unique values? I have a non-dictionary column of id strings like

"0.0.89748-1612413398-472232185"

and

select * where id = '...' performance

is horrible. Just times out after 10 seconds. I assume I need to add a reverse index, the question is should it be dictionary or non-dictionary?

try bloom filter first

Cool interesting, will do 🙂

Also, I notice you are running

select *

, as opposed to

select count(*)

, are you interested in actual rows, or just a count?