kosiew commented on code in PR #25530:
URL: https://github.com/apache/datafusion/pull/25530#discussion_r4205322551
##########
datafusion/physical-plan/src/joins/cross_join.rs:
##########
@@ -966,6 +991,40 @@ mod tests {
");
assert_join_metrics!(metrics, 6);
+ assert_eq!(metric_count(&metrics, "build_input_rows"), 3);
+ assert_eq!(metric_count(&metrics, "input_rows"), 2);
+
+ Ok(())
+ }
+
+ /// `build_input_rows`, `input_rows`, and `output_rows` must count each row
+ /// once when the build and probe sides arrive as several batches.
+ #[tokio::test]
+ async fn join_metrics_count_rows_once_across_batches() -> Result<()> {
+ let task_ctx = Arc::new(TaskContext::default());
+ // 2: left 2+1, right 2. 1: left 3×1, right 2×1.
+ for input_batch_size in [2, 1] {
+ let left = build_table_scan_i32_chunked(
+ ("a1", &vec![1, 2, 3]),
+ ("b1", &vec![4, 5, 6]),
+ ("c1", &vec![7, 8, 9]),
+ input_batch_size,
+ );
+ let right = build_table_scan_i32_chunked(
+ ("a2", &vec![10, 11]),
+ ("b2", &vec![12, 13]),
+ ("c2", &vec![14, 15]),
+ input_batch_size,
+ );
+
+ let (_, batches, metrics) =
+ join_collect(left, right, Arc::clone(&task_ctx)).await?;
+ let num_rows: usize = batches.iter().map(|b| b.num_rows()).sum();
+ assert_eq!(num_rows, 6);
+ assert_join_metrics!(metrics, 6);
+ assert_eq!(metric_count(&metrics, "build_input_rows"), 3);
+ assert_eq!(metric_count(&metrics, "input_rows"), 2);
Review Comment:
Optional: could we also assert `build_input_batches` and `input_batches`,
expecting `(2, 1)` for batch size 2 and `(3, 2)` for batch size 1? That would
ensure future test setup changes do not accidentally remove the multi-batch
coverage while keeping the same row totals.
--
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]
