This message was deleted.
# troubleshootings
Slackbot
09/03/2022, 6:15 PMThis message was deleted.
g
Gian Merlino
09/07/2022, 7:28 AMwhat do you mean by "problem"?
they certainly pose challenges for any distributed DB, although druid's always been good at handling them relative to other options — esp. with features like bitmap indexes and pruning for fast filtering, & various approximate algorithms for ranking, counting, etc
d
Didip Kerabat
09/07/2022, 10:53 PMoh? Maybe my knowledge is old then? I thought the bitmap index for string type could get really big on high cardinality columns?
g
Gian Merlino
09/07/2022, 11:03 PMI suppose it depends on what you mean by "really big" and whether you think it's worth it
Gian Merlino
09/07/2022, 11:04 PMIMO, if you do a lot of
=inGian Merlino
09/07/2022, 11:04 PMif you never do filters like that, and you don't find the index useful, you can disable it on a column by column basis
Gian Merlino
09/07/2022, 11:04 PMcreateBitmapIndex: falsed
Didip Kerabat
09/07/2022, 11:07 PMRelated question, in the old days, circa v0.15.x, broker/historical used to crash if you have very high cardinality string columns, does it no longer happen on groupBy v2 (since groupBy v2 can offload to disk)?
g
Gian Merlino
09/07/2022, 11:13 PMyeah, these days, worst you will get is ResourceLimitExceeded error
🙌 1