In druid 29.0.1, does it always use broadcast join...
# general
p
In druid 29.0.1, does it always use broadcast join, or it uses hash join as well automatically based on table sizes? Or we need to specify?
g
the native engine uses a broadcast hash-join documented here- https://druid.apache.org/docs/latest/querying/query-execution#join
multi-stage-query engine (task based) can use broadcast or sort-merge joins https://druid.apache.org/docs/latest/multi-stage-query/reference/#joins