Any best practice for druid working on 1000+ dimensions?

Hi all,
Yes, we will have more than 1000 dimensions later some time, I wander if there is any best practice working with druid.

let’s say, we have 1000+ dimensions for each row and total row number would be 10M+ for every segment.

Any experiences on this size of dataset is appreciated.



Hi Yanli, I have known folks that have created segments with 1000+ dimensions, but I am not sure if they are in production. Charles might be able to answer more here. Do you have any data where you could experiment with segments of this many dimensions?