martin-g commented on code in PR #18972:
URL: https://github.com/apache/datafusion/pull/18972#discussion_r2571145546
##########
datafusion/core/tests/fuzz_cases/join_fuzz.rs:
##########
@@ -975,6 +975,10 @@ impl JoinFuzzTestCase {
.zip(&hj_formatted_sorted)
.enumerate()
{
+ if !nlj_line.contains("\n") && !hj_line.contains("\n") {
+ // both are empty results
Review Comment:
This check looks wrong.
Both `nlj_formatted_sorted` and `hj_formatted_sorted` are created from
`.lines()` iterator, so their items should not contain new lines.
https://github.com/apache/datafusion/pull/18972/files#diff-f8330244c576c142337679ecbeeb99e9016c2c7c54d7bae492144374e138a0b7R900-R905
##########
datafusion/physical-plan/src/joins/hash_join/stream.rs:
##########
@@ -389,14 +430,22 @@ impl HashJoinStream {
handle_state!(ready!(self.fetch_probe_batch(cx)))
}
HashJoinStreamState::ProcessProbeBatch(_) => {
- let poll = handle_state!(self.process_probe_batch());
- self.join_metrics.baseline.record_poll(poll)
+ handle_state!(self.process_probe_batch())
}
HashJoinStreamState::ExhaustedProbeSide => {
- let poll =
handle_state!(self.process_unmatched_build_batch());
- self.join_metrics.baseline.record_poll(poll)
+ handle_state!(self.process_unmatched_build_batch())
+ }
+ HashJoinStreamState::Completed => {
+ // Flush any remaining buffered data
+ if !self.output_buffer.is_empty() {
+ self.output_buffer.finish_buffered_batch().map_err(
Review Comment:
is `map_err` needed here ?
`impl From<ArrowError> for DataFusionError` (datafusion/common/src/error.rs)
should help to avoid the manual mapping.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
