This is an automated email from the ASF dual-hosted git repository.
lihaopeng pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/doris.git
The following commit(s) were added to refs/heads/master by this push:
new 947397e9994 [Bug](rf) fix rf of in filter cast data as different class
type maybe return wrong result (#39026)
947397e9994 is described below
commit 947397e999429104ce941e13df1d7369f4077160
Author: zhangstar333 <[email protected]>
AuthorDate: Sat Aug 10 18:27:31 2024 +0800
[Bug](rf) fix rf of in filter cast data as different class type maybe
return wrong result (#39026)
two point have changed:
1. in batch_assign function:
const std::string& string_value = column.stringval();
if call **insert(&string_value)**, will cast as string_ref:
reinterpret_cast<const **StringRef***>(data), this maybe error;
```
void insert(const void* data) override {
if (data == nullptr) {
_contains_null = true;
return;
}
const auto* value = reinterpret_cast<const StringRef*>(data);
std::string str_value(value->data, value->size);
_set.insert(str_value);
}
```
2. in batch_copy function, will cast void_value as T*
but the it->get_value() return is StringRef, so need change T as
StringRef
```
template <typename T>
void batch_copy(PInFilter* filter, HybridSetBase::IteratorBase* it,
void (*set_func)(PColumnValue*, const T*)) {
while (it->has_next()) {
const void* void_value = it->get_value();
auto origin_value = reinterpret_cast<const T*>(void_value);
set_func(filter->add_values(), origin_value);
it->next();
}
}
```
---
be/src/exprs/runtime_filter.cpp | 12 ++++++++----
regression-test/data/query_p0/join/test_runtimefilter_2.out | 9 +++++++++
.../suites/query_p0/join/test_runtimefilter_2.groovy | 11 +++++++++++
3 files changed, 28 insertions(+), 4 deletions(-)
diff --git a/be/src/exprs/runtime_filter.cpp b/be/src/exprs/runtime_filter.cpp
index c6fd3338b14..5a241326f90 100644
--- a/be/src/exprs/runtime_filter.cpp
+++ b/be/src/exprs/runtime_filter.cpp
@@ -694,8 +694,10 @@ public:
case TYPE_CHAR:
case TYPE_STRING: {
batch_assign(in_filter, [](std::shared_ptr<HybridSetBase>& set,
PColumnValue& column) {
- const auto& string_val_ref = column.stringval();
- set->insert(&string_val_ref);
+ const std::string& string_value = column.stringval();
+ // string_value is std::string, call insert(data, size)
function in StringSet will not cast as StringRef
+ // so could avoid some cast error at different class object.
+ set->insert((void*)string_value.data(), string_value.size());
});
break;
}
@@ -1630,8 +1632,10 @@ void IRuntimeFilter::to_protobuf(PInFilter* filter) {
case TYPE_CHAR:
case TYPE_VARCHAR:
case TYPE_STRING: {
- batch_copy<std::string>(filter, it, [](PColumnValue* column, const
std::string* value) {
- column->set_stringval(*value);
+ //const void* void_value = it->get_value();
+ //Now the get_value return void* is StringRef
+ batch_copy<StringRef>(filter, it, [](PColumnValue* column, const
StringRef* value) {
+ column->set_stringval(value->to_string());
});
return;
}
diff --git a/regression-test/data/query_p0/join/test_runtimefilter_2.out
b/regression-test/data/query_p0/join/test_runtimefilter_2.out
index d6cc7fc59a0..005406e6793 100644
--- a/regression-test/data/query_p0/join/test_runtimefilter_2.out
+++ b/regression-test/data/query_p0/join/test_runtimefilter_2.out
@@ -2,3 +2,12 @@
-- !select_1 --
aaa
+-- !select_2 --
+aaa
+
+-- !select_3 --
+BSDSAE1018 1 1 true BSDSAE1018 1 true true
+
+-- !select_4 --
+2 3 BSDSAE1018
+
diff --git a/regression-test/suites/query_p0/join/test_runtimefilter_2.groovy
b/regression-test/suites/query_p0/join/test_runtimefilter_2.groovy
index 6e6e57c6c2d..50a61a366b1 100644
--- a/regression-test/suites/query_p0/join/test_runtimefilter_2.groovy
+++ b/regression-test/suites/query_p0/join/test_runtimefilter_2.groovy
@@ -30,4 +30,15 @@
qt_select_1 """
select "aaa" FROM t_ods_tpisyncjpa4_2 tpisyncjpa4
inner join ( SELECT USER_ID, MAX(INTERNAL_CODE)
as INTERNAL_CODE FROM t_ods_tpisyncjpa4_2 WHERE
STATE_ID = '1' GROUP BY USER_ID ) jpa4 on
tpisyncjpa4.USER_ID = jpa4.USER_ID;
"""
+ sql """set runtime_filter_type='IN';"""
+ qt_select_2 """
+ select "aaa" FROM t_ods_tpisyncjpa4_2 tpisyncjpa4
inner join ( SELECT USER_ID, MAX(INTERNAL_CODE)
as INTERNAL_CODE FROM t_ods_tpisyncjpa4_2 WHERE
STATE_ID = '1' GROUP BY USER_ID ) jpa4 on
tpisyncjpa4.USER_ID = jpa4.USER_ID;
+ """
+ qt_select_3 """
+ select *, tpisyncjpp1.POST_ID=jpp1.POST_ID,
tpisyncjpp1.INTERNAL_CODE=jpp1.INTERNAL_CODE from ( select
tpisyncjpp1.POST_ID,tpisyncjpp1.INTERNAL_CODE as INTERNAL_CODE,
tpisyncjpp1.STATE_ID, tpisyncjpp1.STATE_ID ='1' from ( select tpisyncjpa4.*
from t_ods_tpisyncjpa4_2 tpisyncjpa4 inner join [broadcast] (
SELECT USER_ID, MAX(INTERNAL_CODE) as
INTERNAL_CODE FROM t_ods_tpisyncjpa4_2
[...]
+ """
+ qt_select_4 """
+ select DISTINCT tpisyncjpa4.USER_ID as USER_ID,
tpisyncjpa4.USER_NAME as USER_NAME, tpisyncjpp1.POST_ID AS "T4_POST_ID"
FROM t_ods_tpisyncjpa4_2 tpisyncjpa4 cross join [shuffle]
t_ods_tpisyncjpp1_2 tpisyncjpp1 inner join ( SELECT
USER_ID, MAX(INTERNAL_CODE) as INTERNAL_CODE FROM
t_ods_tpisyncjpa4_2 WHERE STATE_ID = '1' GROUP BY
USER_ID )jpa4 on [...]
+ """
+
}
\ No newline at end of file
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
