liukun4515 commented on code in PR #4562:
URL: https://github.com/apache/arrow-datafusion/pull/4562#discussion_r1046762584
##########
datafusion/core/src/physical_plan/planner.rs:
##########
@@ -940,7 +940,16 @@ impl DefaultPhysicalPlanner {
.read()
.get_bool(OPT_PREFER_HASH_JOIN)
.unwrap_or_default();
- if session_state.config.target_partitions() > 1
+ if join_on.is_empty() {
+ // there is no equal join condition, use the nested
loop join
+ // TODO optimize the plan, and use the config of
`target_partitions` and `repartition_joins`
Review Comment:
In the `left out` join type, the distribution of the join is
```
vec![
Distribution::UnspecifiedDistribution,
Distribution::SinglePartition,
]
```
can we partitioned the left side before NLJ exec to accelerate the execution?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
