This is an automated email from the ASF dual-hosted git repository.
yiguolei pushed a commit to branch branch-4.0
in repository https://gitbox.apache.org/repos/asf/doris.git
The following commit(s) were added to refs/heads/branch-4.0 by this push:
new 7c93ad9c4a5 branch-4.0: [fix](local exchange) Do global shuffle for
shuffled join/set operator #60758 (#60763)
7c93ad9c4a5 is described below
commit 7c93ad9c4a5bbb7a95e2049fc66efcd2236eeab2
Author: github-actions[bot]
<41898282+github-actions[bot]@users.noreply.github.com>
AuthorDate: Sat Feb 14 21:39:05 2026 +0800
branch-4.0: [fix](local exchange) Do global shuffle for shuffled join/set
operator #60758 (#60763)
Cherry-picked from #60758
Co-authored-by: Gabriel <[email protected]>
---
be/src/pipeline/exec/hashjoin_build_sink.h | 4 ++++
be/src/pipeline/exec/hashjoin_probe_operator.h | 4 ++++
be/src/pipeline/exec/operator.h | 6 ++++--
be/src/pipeline/exec/partitioned_hash_join_probe_operator.h | 3 +++
be/src/pipeline/exec/partitioned_hash_join_sink_operator.h | 3 +++
be/src/pipeline/exec/set_probe_sink_operator.h | 4 ++++
be/src/pipeline/exec/set_sink_operator.h | 4 ++++
7 files changed, 26 insertions(+), 2 deletions(-)
diff --git a/be/src/pipeline/exec/hashjoin_build_sink.h
b/be/src/pipeline/exec/hashjoin_build_sink.h
index 2bb9407ef5c..f0cf73a82dd 100644
--- a/be/src/pipeline/exec/hashjoin_build_sink.h
+++ b/be/src/pipeline/exec/hashjoin_build_sink.h
@@ -148,6 +148,10 @@ public:
return _join_distribution == TJoinDistributionType::BUCKET_SHUFFLE ||
_join_distribution == TJoinDistributionType::COLOCATE;
}
+ bool followed_by_shuffled_operator() const override {
+ return (is_shuffled_operator() && !is_colocated_operator()) ||
+ _followed_by_shuffled_operator;
+ }
std::vector<bool>& is_null_safe_eq_join() { return _is_null_safe_eq_join; }
bool allow_left_semi_direct_return(RuntimeState* state) const {
diff --git a/be/src/pipeline/exec/hashjoin_probe_operator.h
b/be/src/pipeline/exec/hashjoin_probe_operator.h
index eff32f432ff..6c5232b3720 100644
--- a/be/src/pipeline/exec/hashjoin_probe_operator.h
+++ b/be/src/pipeline/exec/hashjoin_probe_operator.h
@@ -153,6 +153,10 @@ public:
return _join_distribution == TJoinDistributionType::BUCKET_SHUFFLE ||
_join_distribution == TJoinDistributionType::COLOCATE;
}
+ bool followed_by_shuffled_operator() const override {
+ return (is_shuffled_operator() && !is_colocated_operator()) ||
+ _followed_by_shuffled_operator;
+ }
bool need_finalize_variant_column() const { return
_need_finalize_variant_column; }
diff --git a/be/src/pipeline/exec/operator.h b/be/src/pipeline/exec/operator.h
index 5ab3fdf6e7f..875d9ce18ef 100644
--- a/be/src/pipeline/exec/operator.h
+++ b/be/src/pipeline/exec/operator.h
@@ -163,12 +163,14 @@ public:
*/
[[nodiscard]] virtual bool is_shuffled_operator() const { return false; }
/**
- * Return True if this operator is followed by a shuffled operator.
+ * For multiple children's operators, return true if this is a shuffled
operator or this is followed by a shuffled operator (HASH JOIN and SET
OPERATION).
+ *
+ * For single child's operators, return true if this operator is followed
by a shuffled operator.
* For example, in the plan fragment:
* `UNION` -> `SHUFFLED HASH JOIN`
* The `SHUFFLED HASH JOIN` is a shuffled operator so the UNION operator
is followed by a shuffled operator.
*/
- [[nodiscard]] bool followed_by_shuffled_operator() const {
+ [[nodiscard]] virtual bool followed_by_shuffled_operator() const {
return _followed_by_shuffled_operator;
}
/**
diff --git a/be/src/pipeline/exec/partitioned_hash_join_probe_operator.h
b/be/src/pipeline/exec/partitioned_hash_join_probe_operator.h
index 1532acfda33..6419e4676ce 100644
--- a/be/src/pipeline/exec/partitioned_hash_join_probe_operator.h
+++ b/be/src/pipeline/exec/partitioned_hash_join_probe_operator.h
@@ -164,6 +164,9 @@ public:
bool is_colocated_operator() const override {
return _inner_probe_operator->is_colocated_operator();
}
+ bool followed_by_shuffled_operator() const override {
+ return _inner_probe_operator->followed_by_shuffled_operator();
+ }
void update_operator(const TPlanNode& tnode, bool
followed_by_shuffled_operator,
bool require_bucket_distribution) override {
diff --git a/be/src/pipeline/exec/partitioned_hash_join_sink_operator.h
b/be/src/pipeline/exec/partitioned_hash_join_sink_operator.h
index e098d071daa..59eed7aac66 100644
--- a/be/src/pipeline/exec/partitioned_hash_join_sink_operator.h
+++ b/be/src/pipeline/exec/partitioned_hash_join_sink_operator.h
@@ -150,6 +150,9 @@ public:
bool is_shuffled_operator() const override {
return _inner_sink_operator->is_shuffled_operator();
}
+ bool followed_by_shuffled_operator() const override {
+ return _inner_sink_operator->followed_by_shuffled_operator();
+ }
void update_operator(const TPlanNode& tnode, bool
followed_by_shuffled_operator,
bool require_bucket_distribution) override {
diff --git a/be/src/pipeline/exec/set_probe_sink_operator.h
b/be/src/pipeline/exec/set_probe_sink_operator.h
index c0ee8b7731e..896088fce0d 100644
--- a/be/src/pipeline/exec/set_probe_sink_operator.h
+++ b/be/src/pipeline/exec/set_probe_sink_operator.h
@@ -115,6 +115,10 @@ public:
bool is_shuffled_operator() const override { return true; }
bool is_colocated_operator() const override { return _is_colocate; }
+ bool followed_by_shuffled_operator() const override {
+ return (is_shuffled_operator() && !is_colocated_operator()) ||
+ Base::_followed_by_shuffled_operator;
+ }
private:
void _finalize_probe(SetProbeSinkLocalState<is_intersect>& local_state);
diff --git a/be/src/pipeline/exec/set_sink_operator.h
b/be/src/pipeline/exec/set_sink_operator.h
index 513612fb091..11c243ef24d 100644
--- a/be/src/pipeline/exec/set_sink_operator.h
+++ b/be/src/pipeline/exec/set_sink_operator.h
@@ -119,6 +119,10 @@ public:
bool is_shuffled_operator() const override { return true; }
bool is_colocated_operator() const override { return _is_colocate; }
+ bool followed_by_shuffled_operator() const override {
+ return (is_shuffled_operator() && !is_colocated_operator()) ||
+ Base::_followed_by_shuffled_operator;
+ }
private:
template <class HashTableContext, bool is_intersected>
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
