This is an automated email from the ASF dual-hosted git repository.
kxiao pushed a commit to branch branch-2.0
in repository https://gitbox.apache.org/repos/asf/doris.git
The following commit(s) were added to refs/heads/branch-2.0 by this push:
new 104721d1c32 [Fix](predicate pushdown) Common expression not acting on
any slot should not be pushed down (#25901) (#26215)
104721d1c32 is described below
commit 104721d1c322c62fb4b962f44b8ea6af7cf0b3b5
Author: zhiqiang <[email protected]>
AuthorDate: Wed Nov 1 06:29:16 2023 -0500
[Fix](predicate pushdown) Common expression not acting on any slot should
not be pushed down (#25901) (#26215)
---
be/src/vec/exec/scan/vscan_node.cpp | 10 +++-
be/src/vec/exprs/vexpr.h | 16 +++++
.../data/correctness/test_pushdown_common_expr.out | 24 ++++++++
.../correctness/test_pushdown_common_expr.groovy | 69 +++++++++++++++++++++-
4 files changed, 116 insertions(+), 3 deletions(-)
diff --git a/be/src/vec/exec/scan/vscan_node.cpp
b/be/src/vec/exec/scan/vscan_node.cpp
index 68e66231300..e0c954042d1 100644
--- a/be/src/vec/exec/scan/vscan_node.cpp
+++ b/be/src/vec/exec/scan/vscan_node.cpp
@@ -17,6 +17,7 @@
#include "vec/exec/scan/vscan_node.h"
+#include <gen_cpp/Exprs_types.h>
#include <gen_cpp/Metrics_types.h>
#include <gen_cpp/Opcodes_types.h>
#include <gen_cpp/PaloInternalService_types.h>
@@ -414,12 +415,16 @@ Status VScanNode::_normalize_conjuncts() {
RETURN_IF_ERROR(_normalize_predicate(conjunct->root(),
conjunct.get(), new_root));
if (new_root) {
conjunct->set_root(new_root);
- if (_should_push_down_common_expr()) {
+ if (_should_push_down_common_expr() &&
+ VExpr::is_acting_on_a_slot(conjunct->root())) {
+ // We need to make sure conjunct is acting on a slot
before push it down.
+ // Or it will not be executed by
SegmentIterator::_vec_init_lazy_materialization
_common_expr_ctxs_push_down.emplace_back(conjunct);
it = _conjuncts.erase(it);
continue;
}
- } else { // All conjuncts are pushed down as predicate column
+ } else {
+ // Whole conjunct is pushed down as predicate column
_stale_expr_ctxs.emplace_back(conjunct);
it = _conjuncts.erase(it);
continue;
@@ -427,6 +432,7 @@ Status VScanNode::_normalize_conjuncts() {
}
++it;
}
+
for (auto& it : _slot_id_to_value_range) {
std::visit(
[&](auto&& range) {
diff --git a/be/src/vec/exprs/vexpr.h b/be/src/vec/exprs/vexpr.h
index 6b632eeee57..ec1c8affc7e 100644
--- a/be/src/vec/exprs/vexpr.h
+++ b/be/src/vec/exprs/vexpr.h
@@ -74,6 +74,22 @@ public:
return block->columns() - 1;
}
+ static bool is_acting_on_a_slot(const VExprSPtr& expr) {
+ const auto& children = expr->children();
+ const size_t children_size = children.size();
+
+ for (size_t i = 0; i < children_size; ++i) {
+ // If any child expr acts on a column slot
+ // return true immediately.
+ if (is_acting_on_a_slot(children[i])) {
+ return true;
+ }
+ }
+
+ // This is a leaf expression.
+ return expr->node_type() == TExprNodeType::SLOT_REF;
+ }
+
VExpr(const TExprNode& node);
VExpr(const VExpr& vexpr);
VExpr(const TypeDescriptor& type, bool is_slotref, bool is_nullable);
diff --git a/regression-test/data/correctness/test_pushdown_common_expr.out
b/regression-test/data/correctness/test_pushdown_common_expr.out
index e50a453c5c3..8da380e841b 100644
--- a/regression-test/data/correctness/test_pushdown_common_expr.out
+++ b/regression-test/data/correctness/test_pushdown_common_expr.out
@@ -22,6 +22,18 @@
64 g gg
8 d dd
+-- !5 --
+256 i ii
+
+-- !6 --
+
+-- !7 --
+
+-- !8 --
+1024 k kk
+
+-- !9 --
+
-- !1 --
1 a aa
128 h hh
@@ -47,3 +59,15 @@
64 g gg
8 d dd
+-- !5 --
+256 i ii
+
+-- !6 --
+
+-- !7 --
+
+-- !8 --
+1024 k kk
+
+-- !9 --
+
diff --git
a/regression-test/suites/correctness/test_pushdown_common_expr.groovy
b/regression-test/suites/correctness/test_pushdown_common_expr.groovy
index 8819288e103..2395f057b1d 100644
--- a/regression-test/suites/correctness/test_pushdown_common_expr.groovy
+++ b/regression-test/suites/correctness/test_pushdown_common_expr.groovy
@@ -65,6 +65,44 @@ suite("test_pushdown_common_expr") {
SELECT * FROM t_pushdown_common_expr WHERE c1 < 300 OR UPPER(c2)="F"
OR c3 LIKE "%f%";
"""
+ // One conjunct. No push down.
+ order_qt_5 """
+ SELECT * FROM t_pushdown_common_expr WHERE random() > 1 OR
uuid_numeric() = 'A' OR c1 = 256
+ """
+
+ // One conjunct. No push down.
+ order_qt_6 """
+ SELECT * FROM t_pushdown_common_expr WHERE random() > 1
+ """
+
+ // Two conjuncts.
+ // random() > 1 is executed by scan node, c1 > 0 is pushed down and
executed by segment iterator
+ order_qt_7 """
+ SELECT * FROM t_pushdown_common_expr WHERE random() > 1 AND c1 > 0
+ """
+
+ // One conjunct. No push down.
+ order_qt_8 """
+ SELECT * FROM t_pushdown_common_expr WHERE random() > 1 OR c1 > 1020
+ """
+
+ // t1: 512 & 1024
+ // t2: all
+ // join gets nothing
+ order_qt_9 """
+ SELECT *
+ FROM (
+ SELECT c1
+ FROM t_pushdown_common_expr
+ WHERE c1 >= 512 or random() > 1
+ ) AS t1
+ JOIN (
+ SELECT c1
+ FROM t_pushdown_common_expr
+ WHERE random() = 0
+ ) AS t2 ON t1.c1 = t2.c1
+ """
+
sql """set enable_common_expr_pushdown=false"""
order_qt_1 """
@@ -83,4 +121,33 @@ suite("test_pushdown_common_expr") {
SELECT * FROM t_pushdown_common_expr WHERE c1 < 300 OR UPPER(c2)="K"
OR c3 LIKE "%k%";
"""
-}
+ order_qt_5 """
+ SELECT * FROM t_pushdown_common_expr WHERE random() > 1 OR
uuid_numeric() = 'A' OR c1 = 256
+ """
+
+ order_qt_6 """
+ SELECT * FROM t_pushdown_common_expr WHERE random() > 1
+ """
+
+ order_qt_7 """
+ SELECT * FROM t_pushdown_common_expr WHERE random() > 1 AND c1 > 0
+ """
+
+ order_qt_8 """
+ SELECT * FROM t_pushdown_common_expr WHERE random() > 1 OR c1 > 1020
+ """
+
+ order_qt_9 """
+ SELECT *
+ FROM (
+ SELECT c1
+ FROM t_pushdown_common_expr
+ WHERE c1 >= 512 or random() > 1
+ ) AS t1
+ JOIN (
+ SELECT c1
+ FROM t_pushdown_common_expr
+ WHERE random() = 0
+ ) AS t2 ON t1.c1 = t2.c1
+ """
+}
\ No newline at end of file
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
