This is an automated email from the ASF dual-hosted git repository.
yiguolei pushed a commit to branch branch-4.1
in repository https://gitbox.apache.org/repos/asf/doris.git
The following commit(s) were added to refs/heads/branch-4.1 by this push:
new ea18ba1f84b branch-4.1: [fix](be) Defer Bloom allocation for IN
runtime filters #68763 (#68789)
ea18ba1f84b is described below
commit ea18ba1f84b49be09d19f0384d2305373b964c14
Author: github-actions[bot]
<41898282+github-actions[bot]@users.noreply.github.com>
AuthorDate: Fri Oct 9 09:26:35 2026 +0800
branch-4.1: [fix](be) Defer Bloom allocation for IN runtime filters #68763
(#68789)
Cherry-picked from #68763
Co-authored-by: Mryange <[email protected]>
---
.../exec/runtime_filter/runtime_filter_wrapper.cpp | 2 +-
.../runtime_filter/runtime_filter_wrapper_test.cpp | 42 ++++++++++++++++++++++
2 files changed, 43 insertions(+), 1 deletion(-)
diff --git a/be/src/exec/runtime_filter/runtime_filter_wrapper.cpp
b/be/src/exec/runtime_filter/runtime_filter_wrapper.cpp
index 4b0a8ca8090..330bcb6895e 100644
--- a/be/src/exec/runtime_filter/runtime_filter_wrapper.cpp
+++ b/be/src/exec/runtime_filter/runtime_filter_wrapper.cpp
@@ -75,7 +75,7 @@ Status RuntimeFilterWrapper::init(const size_t real_size) {
if (get_real_type() == RuntimeFilterType::IN_FILTER && real_size >
_max_in_num) {
set_state(RuntimeFilterWrapper::State::DISABLED, "reach max in num");
}
- if (_bloom_filter_func) {
+ if (get_real_type() == RuntimeFilterType::BLOOM_FILTER) {
RETURN_IF_ERROR(_bloom_filter_func->init_with_fixed_length(real_size));
}
return Status::OK();
diff --git a/be/test/exec/runtime_filter/runtime_filter_wrapper_test.cpp
b/be/test/exec/runtime_filter/runtime_filter_wrapper_test.cpp
index d5b2f9d9641..44cde989ab0 100644
--- a/be/test/exec/runtime_filter/runtime_filter_wrapper_test.cpp
+++ b/be/test/exec/runtime_filter/runtime_filter_wrapper_test.cpp
@@ -749,6 +749,46 @@ TEST_F(RuntimeFilterWrapperTest, TestMinMax) {
}
}
+TEST_F(RuntimeFilterWrapperTest, TestInOrBloomLazyInitialization) {
+ for (bool build_by_runtime_size : {false, true}) {
+ RuntimeFilterParams params {.filter_id = 0,
+ .filter_type =
RuntimeFilterType::IN_OR_BLOOM_FILTER,
+ .column_return_type = TYPE_INT,
+ .null_aware = false,
+ .max_in_num = 16,
+ .runtime_bloom_filter_min_size = 64,
+ .runtime_bloom_filter_max_size = 128,
+ .bloom_filter_size = 64,
+ .build_bf_by_runtime_size =
build_by_runtime_size,
+ .bloom_filter_size_calculated_by_ndv =
false};
+ for (size_t runtime_size : {0, 1, 16, 17}) {
+ SCOPED_TRACE(runtime_size);
+ SCOPED_TRACE(build_by_runtime_size);
+ RuntimeFilterWrapper wrapper(¶ms);
+ ASSERT_NE(wrapper.bloom_filter_func(), nullptr);
+ ASSERT_EQ(wrapper.bloom_filter_func()->get_size(), 0);
+ ASSERT_TRUE(wrapper.init(runtime_size).ok());
+ if (runtime_size <= static_cast<size_t>(params.max_in_num)) {
+ EXPECT_EQ(wrapper.get_real_type(),
RuntimeFilterType::IN_FILTER);
+ EXPECT_EQ(wrapper.bloom_filter_func()->get_size(), 0);
+ EXPECT_EQ(wrapper.bloom_filter_func()->_bloom_filter_alloced,
0);
+ EXPECT_EQ(wrapper.bloom_filter_func()->_bloom_filter, nullptr);
+ PMergeFilterRequest request;
+
ASSERT_TRUE(wrapper.to_protobuf(request.mutable_in_filter()).ok());
+ } else {
+ EXPECT_EQ(wrapper.get_real_type(),
RuntimeFilterType::BLOOM_FILTER);
+ EXPECT_GT(wrapper.bloom_filter_func()->get_size(), 0);
+ auto col = ColumnHelper::create_column<DataTypeInt32>({1, 2,
3});
+ ASSERT_TRUE(wrapper.insert(col, 0).ok());
+ std::vector<uint8_t> matches(3);
+ wrapper.bloom_filter_func()->find_fixed_len(col,
matches.data());
+ EXPECT_TRUE(std::all_of(matches.begin(), matches.end(),
+ [](uint8_t match) { return match != 0;
}));
+ }
+ }
+ }
+}
+
TEST_F(RuntimeFilterWrapperTest, TestInOrBloom) {
std::vector<int> data_vector(10);
std::iota(data_vector.begin(), data_vector.end(), 0);
@@ -817,6 +857,7 @@ TEST_F(RuntimeFilterWrapperTest, TestInOrBloom) {
// Init (keep in filter)
EXPECT_TRUE(wrapper->init(runtime_size).ok());
EXPECT_EQ(wrapper->get_real_type(), RuntimeFilterType::IN_FILTER);
+ EXPECT_EQ(wrapper->bloom_filter_func()->get_size(), 0);
EXPECT_EQ(wrapper->get_state(), RuntimeFilterWrapper::State::UNINITED);
// Insert
auto col = ColumnHelper::create_column<DataType>(data_vector);
@@ -875,6 +916,7 @@ TEST_F(RuntimeFilterWrapperTest, TestInOrBloom) {
EXPECT_TRUE(wrapper->merge(new_wrapper.get()).ok());
EXPECT_EQ(wrapper->hybrid_set()->size(), col->size() * 2);
EXPECT_EQ(wrapper->get_real_type(), RuntimeFilterType::IN_FILTER);
+ EXPECT_EQ(wrapper->bloom_filter_func()->get_size(), 0);
}
{
// In + In -> Bloom
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
