This is an automated email from the ASF dual-hosted git repository.
yiguolei 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 c4341d3d43 [fix](like)prevent null pointer by unimplemented like_vec
functions (#12910)
c4341d3d43 is described below
commit c4341d3d43c82387a7f065e2dfd69fcd3822c003
Author: starocean999 <[email protected]>
AuthorDate: Tue Sep 27 10:02:10 2022 +0800
[fix](like)prevent null pointer by unimplemented like_vec functions (#12910)
* [fix](like)prevent null pointer by unimplemented like_vec functions
* fix pushed like predicate on dict encoded column bug
---
be/src/olap/like_column_predicate.cpp | 81 ++++++++++++++++++++++++----------
be/src/olap/like_column_predicate.h | 62 +++++++++++++++++---------
be/src/vec/columns/column_dictionary.h | 8 ++++
3 files changed, 107 insertions(+), 44 deletions(-)
diff --git a/be/src/olap/like_column_predicate.cpp
b/be/src/olap/like_column_predicate.cpp
index 56729c1c45..e0c9851f26 100644
--- a/be/src/olap/like_column_predicate.cpp
+++ b/be/src/olap/like_column_predicate.cpp
@@ -88,7 +88,7 @@ uint16_t LikeColumnPredicate<is_vectorized>::evaluate(const
vectorized::IColumn&
continue;
}
- StringValue cell_value =
nested_col_ptr->get_value(data_array[idx]);
+ StringValue cell_value =
nested_col_ptr->get_shrink_value(data_array[idx]);
unsigned char flag = 0;
(_state->function)(const_cast<vectorized::LikeSearchState*>(&_like_state),
cell_value, pattern, &flag);
@@ -115,45 +115,78 @@ uint16_t
LikeColumnPredicate<is_vectorized>::evaluate(const vectorized::IColumn&
}
} else {
if (column.is_column_dictionary()) {
- if (LIKELY(_like_state.search_string_sv.len > 0)) {
- auto* nested_col_ptr = vectorized::check_and_get_column<
-
vectorized::ColumnDictionary<vectorized::Int32>>(column);
- auto& data_array = nested_col_ptr->get_data();
- StringValue values[size];
- unsigned char flags[size];
- for (uint16_t i = 0; i != size; i++) {
- values[i] =
nested_col_ptr->get_value(data_array[sel[i]]);
- }
- (_state->function_vec_dict)(
-
const_cast<vectorized::LikeSearchState*>(&_like_state), pattern, values,
- size, flags);
+ if (_state->function_vec_dict) {
+ if (LIKELY(_like_state.search_string_sv.len > 0)) {
+ auto* nested_col_ptr =
vectorized::check_and_get_column<
+
vectorized::ColumnDictionary<vectorized::Int32>>(column);
+ auto& data_array = nested_col_ptr->get_data();
+ StringValue values[size];
+ unsigned char flags[size];
+ for (uint16_t i = 0; i != size; i++) {
+ values[i] =
nested_col_ptr->get_shrink_value(data_array[sel[i]]);
+ }
+ (_state->function_vec_dict)(
+
const_cast<vectorized::LikeSearchState*>(&_like_state), pattern,
+ values, size, flags);
- for (uint16_t i = 0; i != size; i++) {
- uint16_t idx = sel[i];
- sel[new_size] = idx;
- new_size += _opposite ^ flags[i];
+ for (uint16_t i = 0; i != size; i++) {
+ uint16_t idx = sel[i];
+ sel[new_size] = idx;
+ new_size += _opposite ^ flags[i];
+ }
+ } else {
+ for (uint16_t i = 0; i != size; i++) {
+ uint16_t idx = sel[i];
+ sel[new_size] = idx;
+ new_size += _opposite ^ true;
+ }
}
} else {
+ auto* nested_col_ptr = vectorized::check_and_get_column<
+
vectorized::ColumnDictionary<vectorized::Int32>>(column);
+ auto& data_array = nested_col_ptr->get_data();
for (uint16_t i = 0; i != size; i++) {
uint16_t idx = sel[i];
sel[new_size] = idx;
- new_size += _opposite ^ true;
+ StringValue cell_value =
nested_col_ptr->get_value(data_array[idx]);
+ unsigned char flag = 0;
+
(_state->function)(const_cast<vectorized::LikeSearchState*>(&_like_state),
+ cell_value, pattern, &flag);
+ new_size += _opposite ^ flag;
}
}
} else {
- if (LIKELY(_like_state.search_string_sv.len > 0)) {
+ if (_state->function_vec) {
+ if (LIKELY(_like_state.search_string_sv.len > 0)) {
+ auto* data_array =
+ vectorized::check_and_get_column<
+
vectorized::PredicateColumnType<TYPE_STRING>>(column)
+ ->get_data()
+ .data();
+
+ (_state->function_vec)(
+
const_cast<vectorized::LikeSearchState*>(&_like_state), pattern,
+ data_array, sel, size, _opposite, &new_size);
+ } else {
+ for (uint16_t i = 0; i < size; i++) {
+ uint16_t idx = sel[i];
+ sel[new_size] = idx;
+ new_size += _opposite ^ true;
+ }
+ }
+ } else {
auto* data_array = vectorized::check_and_get_column<
vectorized::PredicateColumnType<TYPE_STRING>>(column)
->get_data()
.data();
-
(_state->function_vec)(const_cast<vectorized::LikeSearchState*>(&_like_state),
- pattern, data_array, sel, size,
_opposite, &new_size);
- } else {
- for (uint16_t i = 0; i < size; i++) {
+ for (uint16_t i = 0; i != size; i++) {
uint16_t idx = sel[i];
sel[new_size] = idx;
- new_size += _opposite ^ true;
+ unsigned char flag = 0;
+
(_state->function)(const_cast<vectorized::LikeSearchState*>(&_like_state),
+ data_array[idx], pattern, &flag);
+ new_size += _opposite ^ flag;
}
}
}
diff --git a/be/src/olap/like_column_predicate.h
b/be/src/olap/like_column_predicate.h
index 303e176461..662f4ee59b 100644
--- a/be/src/olap/like_column_predicate.h
+++ b/be/src/olap/like_column_predicate.h
@@ -98,7 +98,7 @@ private:
continue;
}
- StringValue cell_value =
nested_col_ptr->get_value(data_array[i]);
+ StringValue cell_value =
nested_col_ptr->get_shrink_value(data_array[i]);
if constexpr (is_and) {
unsigned char flag = 0;
(_state->function)(
@@ -118,31 +118,53 @@ private:
}
} else {
if (column.is_column_dictionary()) {
- if (LIKELY(_like_state.search_string_sv.len > 0)) {
- auto* nested_col_ptr =
vectorized::check_and_get_column<
-
vectorized::ColumnDictionary<vectorized::Int32>>(column);
- auto& data_array = nested_col_ptr->get_data();
- StringValue values[size];
- unsigned char temp_flags[size];
- for (uint16_t i = 0; i != size; i++) {
- values[i] =
nested_col_ptr->get_value(data_array[i]);
- }
- (_state->function_vec_dict)(
-
const_cast<vectorized::LikeSearchState*>(&_like_state), pattern,
- values, size, temp_flags);
- for (uint16_t i = 0; i < size; i++) {
- if constexpr (is_and) {
- flags[i] &= _opposite ^ temp_flags[i];
- } else {
- flags[i] = _opposite ^ temp_flags[i];
+ if (_state->function_vec_dict) {
+ if (LIKELY(_like_state.search_string_sv.len > 0)) {
+ auto* nested_col_ptr =
vectorized::check_and_get_column<
+
vectorized::ColumnDictionary<vectorized::Int32>>(column);
+ auto& data_array = nested_col_ptr->get_data();
+ StringValue values[size];
+ unsigned char temp_flags[size];
+ for (uint16_t i = 0; i != size; i++) {
+ values[i] =
nested_col_ptr->get_shrink_value(data_array[i]);
+ }
+ (_state->function_vec_dict)(
+
const_cast<vectorized::LikeSearchState*>(&_like_state), pattern,
+ values, size, temp_flags);
+ for (uint16_t i = 0; i < size; i++) {
+ if constexpr (is_and) {
+ flags[i] &= _opposite ^ temp_flags[i];
+ } else {
+ flags[i] = _opposite ^ temp_flags[i];
+ }
+ }
+ } else {
+ for (uint16_t i = 0; i < size; i++) {
+ if constexpr (is_and) {
+ flags[i] &= _opposite ^ true;
+ } else {
+ flags[i] = _opposite ^ true;
+ }
}
}
} else {
+ auto* nested_col_ptr =
vectorized::check_and_get_column<
+
vectorized::ColumnDictionary<vectorized::Int32>>(column);
+ auto& data_array = nested_col_ptr->get_data();
for (uint16_t i = 0; i < size; i++) {
+ StringValue cell_value =
nested_col_ptr->get_value(data_array[i]);
if constexpr (is_and) {
- flags[i] &= _opposite ^ true;
+ unsigned char flag = 0;
+ (_state->function)(
+
const_cast<vectorized::LikeSearchState*>(&_like_state),
+ cell_value, pattern, &flag);
+ flags[i] &= _opposite ^ flag;
} else {
- flags[i] = _opposite ^ true;
+ unsigned char flag = 0;
+ (_state->function)(
+
const_cast<vectorized::LikeSearchState*>(&_like_state),
+ cell_value, pattern, &flag);
+ flags[i] = _opposite ^ flag;
}
}
}
diff --git a/be/src/vec/columns/column_dictionary.h
b/be/src/vec/columns/column_dictionary.h
index 2c09d4fb7c..7aa578d665 100644
--- a/be/src/vec/columns/column_dictionary.h
+++ b/be/src/vec/columns/column_dictionary.h
@@ -281,6 +281,14 @@ public:
inline const StringValue& get_value(value_type code) const { return
_dict.get_value(code); }
+ inline StringValue get_shrink_value(value_type code) const {
+ StringValue result = _dict.get_value(code);
+ if (_type == OLAP_FIELD_TYPE_CHAR) {
+ result.len = strnlen(result.ptr, result.len);
+ }
+ return result;
+ }
+
class Dictionary {
public:
Dictionary() : _dict_data(new DictContainer()), _total_str_len(0) {};
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
