github-actions[bot] commented on code in PR #26089:
URL: https://github.com/apache/doris/pull/26089#discussion_r1377169999
##########
be/src/vec/columns/column_string.cpp:
##########
@@ -161,6 +161,43 @@ void ColumnString::insert_indices_from(const IColumn& src,
const int* indices_be
}
}
+void ColumnString::insert_indices_from_join(const IColumn& src, const
uint32_t* indices_begin,
Review Comment:
warning: method 'insert_indices_from_join' can be made static
[readability-convert-member-functions-to-static]
be/src/vec/columns/column_string.h:477:
```diff
- void insert_indices_from_join(const IColumn& src, const uint32_t*
indices_begin,
+ static void insert_indices_from_join(const IColumn& src, const
uint32_t* indices_begin,
```
##########
be/src/exprs/runtime_filter.cpp:
##########
@@ -508,24 +476,33 @@ class RuntimePredicateWrapper {
}
}
- void insert_batch(const vectorized::ColumnPtr column, const
std::vector<int>& rows) {
+ void insert_batch(const vectorized::ColumnPtr& column, size_t start) {
if (get_real_type() == RuntimeFilterType::BITMAP_FILTER) {
- bitmap_filter_insert_batch(column, rows);
- } else if (IRuntimeFilter::enable_use_batch(_be_exec_version > 0,
_column_return_type)) {
- insert_fixed_len(column->get_raw_data().data, rows.data(),
rows.size());
+ bitmap_filter_insert_batch(column, start);
} else {
- for (int index : rows) {
- insert(column->get_data_at(index));
- }
+ insert_fixed_len(column, start);
}
}
- void bitmap_filter_insert_batch(const vectorized::ColumnPtr column,
- const std::vector<int>& rows) {
+ void bitmap_filter_insert_batch(const vectorized::ColumnPtr column, size_t
start) {
Review Comment:
warning: method 'bitmap_filter_insert_batch' can be made static
[readability-convert-member-functions-to-static]
```suggestion
static void bitmap_filter_insert_batch(const vectorized::ColumnPtr
column, size_t start) {
```
##########
be/src/vec/columns/column_string.cpp:
##########
@@ -161,6 +161,43 @@
}
}
+void ColumnString::insert_indices_from_join(const IColumn& src, const
uint32_t* indices_begin,
+ const uint32_t* indices_end) {
+ const ColumnString& src_str = assert_cast<const ColumnString&>(src);
+ auto src_offset_data = src_str.offsets.data();
+
+ auto old_char_size = chars.size();
+ size_t total_chars_size = old_char_size;
+
+ auto dst_offsets_pos = offsets.size();
+ offsets.resize(offsets.size() + indices_end - indices_begin);
+ auto* dst_offsets_data = offsets.data();
+
+ for (auto x = indices_begin; x != indices_end; ++x) {
+ if (*x != 0) {
+ total_chars_size += src_offset_data[*x] - src_offset_data[*x - 1];
+ }
+ dst_offsets_data[dst_offsets_pos++] = total_chars_size;
+ }
+ check_chars_length(total_chars_size, offsets.size());
+
+ chars.resize(total_chars_size);
+
+ auto* src_data_ptr = src_str.chars.data();
+ auto* dst_data_ptr = chars.data();
+
+ size_t dst_chars_pos = old_char_size;
+ for (auto x = indices_begin; x != indices_end; ++x) {
Review Comment:
warning: 'auto x' can be declared as 'const auto *x'
[readability-qualified-auto]
```suggestion
for (const auto *x = indices_begin; x != indices_end; ++x) {
```
##########
be/src/vec/common/hash_table/hash_map.h:
##########
@@ -20,6 +20,10 @@
#pragma once
+#include <gen_cpp/PlanNodes_types.h>
Review Comment:
warning: 'gen_cpp/PlanNodes_types.h' file not found [clang-diagnostic-error]
```cpp
#include <gen_cpp/PlanNodes_types.h>
^
```
##########
be/src/vec/common/hash_table/hash_map.h:
##########
@@ -193,10 +197,214 @@
bool has_null_key_data() const { return false; }
};
+template <typename Key, typename Cell, typename Hash = DefaultHash<Key>,
+ typename Grower = HashTableGrower<>, typename Allocator =
HashTableAllocator>
+class JoinHashMapTable : public HashMapTable<Key, Cell, Hash, Grower,
Allocator> {
+public:
+ using Self = JoinHashMapTable;
+ using Base = HashMapTable<Key, Cell, Hash, Grower, Allocator>;
+
+ using key_type = Key;
+ using value_type = typename Cell::value_type;
+ using mapped_type = typename Cell::Mapped;
+
+ using LookupResult = typename Base::LookupResult;
+
+ using HashMapTable<Key, Cell, Hash, Grower, Allocator>::HashMapTable;
+
+ static uint32_t calc_bucket_size(size_t num_elem) {
+ size_t expect_bucket_size = static_cast<size_t>(num_elem) + (num_elem
- 1) / 7;
+ return phmap::priv::NormalizeCapacity(expect_bucket_size) + 1;
+ }
+
+ template <int JoinOpType>
+ void prepare_build(size_t num_elem, int batch_size) {
+ max_batch_size = batch_size;
+ bucket_size = calc_bucket_size(num_elem + 1);
+ first.resize(bucket_size, 0);
+ next.resize(num_elem);
+
+ if constexpr (JoinOpType == doris::TJoinOp::FULL_OUTER_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_OUTER_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_ANTI_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_SEMI_JOIN) {
+ visited.resize(num_elem, 0);
+ }
+ }
+
+ uint32_t get_bucket_size() const { return bucket_size; }
+
+ size_t size() const { return next.size(); }
+
+ void build(const Key* __restrict keys, const uint32_t* __restrict
bucket_nums,
+ size_t num_elem) {
+ build_keys = keys;
+ for (size_t i = 1; i < num_elem; i++) {
+ uint32_t bucket_num = bucket_nums[i];
+ next[i] = first[bucket_num];
+ first[bucket_num] = i;
+ }
+ }
+
+ template <int JoinOpType>
+ auto find_batch(const Key* __restrict keys, const uint32_t* __restrict
bucket_nums,
+ int probe_idx, uint32_t build_idx, int probe_rows,
+ uint32_t* __restrict probe_idxs, uint32_t* __restrict
build_idxs) {
+ if constexpr (JoinOpType == doris::TJoinOp::INNER_JOIN ||
+ JoinOpType == doris::TJoinOp::FULL_OUTER_JOIN ||
+ JoinOpType == doris::TJoinOp::LEFT_OUTER_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_OUTER_JOIN) {
+ return _find_batch_inner_outer_join<JoinOpType>(keys, bucket_nums,
probe_idx, build_idx,
+ probe_rows,
probe_idxs, build_idxs);
+ }
+ if constexpr (JoinOpType == doris::TJoinOp::LEFT_ANTI_JOIN ||
+ JoinOpType == doris::TJoinOp::LEFT_SEMI_JOIN) {
+ return _find_batch_left_semi_anti<JoinOpType>(keys, bucket_nums,
probe_idx, probe_rows,
+ probe_idxs);
+ }
+ if constexpr (JoinOpType == doris::TJoinOp::RIGHT_ANTI_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_SEMI_JOIN) {
+ return _find_batch_right_semi_anti(keys, bucket_nums, probe_idx,
probe_rows);
+ }
+ return std::tuple {0, 0u, 0};
Review Comment:
warning: integer literal has suffix 'u', which is not uppercase
[readability-uppercase-literal-suffix]
```suggestion
return std::tuple {0, 0U, 0};
```
##########
be/src/vec/columns/column_string.cpp:
##########
@@ -161,6 +161,43 @@
}
}
+void ColumnString::insert_indices_from_join(const IColumn& src, const
uint32_t* indices_begin,
+ const uint32_t* indices_end) {
+ const ColumnString& src_str = assert_cast<const ColumnString&>(src);
+ auto src_offset_data = src_str.offsets.data();
+
+ auto old_char_size = chars.size();
+ size_t total_chars_size = old_char_size;
+
+ auto dst_offsets_pos = offsets.size();
+ offsets.resize(offsets.size() + indices_end - indices_begin);
+ auto* dst_offsets_data = offsets.data();
+
+ for (auto x = indices_begin; x != indices_end; ++x) {
Review Comment:
warning: 'auto x' can be declared as 'const auto *x'
[readability-qualified-auto]
```suggestion
for (const auto *x = indices_begin; x != indices_end; ++x) {
```
##########
be/src/vec/common/hash_table/hash_map.h:
##########
@@ -193,10 +197,214 @@
bool has_null_key_data() const { return false; }
};
+template <typename Key, typename Cell, typename Hash = DefaultHash<Key>,
+ typename Grower = HashTableGrower<>, typename Allocator =
HashTableAllocator>
+class JoinHashMapTable : public HashMapTable<Key, Cell, Hash, Grower,
Allocator> {
+public:
+ using Self = JoinHashMapTable;
+ using Base = HashMapTable<Key, Cell, Hash, Grower, Allocator>;
+
+ using key_type = Key;
+ using value_type = typename Cell::value_type;
+ using mapped_type = typename Cell::Mapped;
+
+ using LookupResult = typename Base::LookupResult;
+
+ using HashMapTable<Key, Cell, Hash, Grower, Allocator>::HashMapTable;
+
+ static uint32_t calc_bucket_size(size_t num_elem) {
+ size_t expect_bucket_size = static_cast<size_t>(num_elem) + (num_elem
- 1) / 7;
+ return phmap::priv::NormalizeCapacity(expect_bucket_size) + 1;
+ }
+
+ template <int JoinOpType>
+ void prepare_build(size_t num_elem, int batch_size) {
+ max_batch_size = batch_size;
+ bucket_size = calc_bucket_size(num_elem + 1);
+ first.resize(bucket_size, 0);
+ next.resize(num_elem);
+
+ if constexpr (JoinOpType == doris::TJoinOp::FULL_OUTER_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_OUTER_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_ANTI_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_SEMI_JOIN) {
+ visited.resize(num_elem, 0);
+ }
+ }
+
+ uint32_t get_bucket_size() const { return bucket_size; }
+
+ size_t size() const { return next.size(); }
+
+ void build(const Key* __restrict keys, const uint32_t* __restrict
bucket_nums,
+ size_t num_elem) {
+ build_keys = keys;
+ for (size_t i = 1; i < num_elem; i++) {
+ uint32_t bucket_num = bucket_nums[i];
+ next[i] = first[bucket_num];
+ first[bucket_num] = i;
+ }
+ }
+
+ template <int JoinOpType>
+ auto find_batch(const Key* __restrict keys, const uint32_t* __restrict
bucket_nums,
+ int probe_idx, uint32_t build_idx, int probe_rows,
+ uint32_t* __restrict probe_idxs, uint32_t* __restrict
build_idxs) {
+ if constexpr (JoinOpType == doris::TJoinOp::INNER_JOIN ||
+ JoinOpType == doris::TJoinOp::FULL_OUTER_JOIN ||
+ JoinOpType == doris::TJoinOp::LEFT_OUTER_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_OUTER_JOIN) {
+ return _find_batch_inner_outer_join<JoinOpType>(keys, bucket_nums,
probe_idx, build_idx,
+ probe_rows,
probe_idxs, build_idxs);
+ }
+ if constexpr (JoinOpType == doris::TJoinOp::LEFT_ANTI_JOIN ||
+ JoinOpType == doris::TJoinOp::LEFT_SEMI_JOIN) {
+ return _find_batch_left_semi_anti<JoinOpType>(keys, bucket_nums,
probe_idx, probe_rows,
+ probe_idxs);
+ }
+ if constexpr (JoinOpType == doris::TJoinOp::RIGHT_ANTI_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_SEMI_JOIN) {
+ return _find_batch_right_semi_anti(keys, bucket_nums, probe_idx,
probe_rows);
+ }
+ return std::tuple {0, 0u, 0};
+ }
+
+ template <int JoinOpType>
+ bool iterate_map(std::vector<uint32_t>& build_idxs) const {
+ const auto batch_size = max_batch_size;
+ const auto elem_num = visited.size();
+ int count = 0;
+ build_idxs.resize(batch_size);
+
+ while (count < batch_size && iter_idx < elem_num) {
+ const auto matched = visited[iter_idx];
+ build_idxs[count] = iter_idx;
+ if constexpr (JoinOpType == doris::TJoinOp::RIGHT_ANTI_JOIN) {
+ count += !matched;
+ } else {
+ count += matched;
+ }
+ iter_idx++;
+ }
+
+ build_idxs.resize(count);
+ return iter_idx >= elem_num;
+ }
+
+private:
+ auto _find_batch_right_semi_anti(const Key* __restrict keys,
+ const uint32_t* __restrict bucket_nums,
int probe_idx,
+ int probe_rows) {
+ while (probe_idx < probe_rows) {
+ auto build_idx = first[bucket_nums[probe_idx]];
+
+ while (build_idx) {
+ if (keys[probe_idx] == build_keys[build_idx]) {
+ visited[build_idx] = 1;
+ }
+ build_idx = next[build_idx];
+ }
+ probe_idx++;
+ }
+ return std::tuple {probe_idx, 0u, 0};
Review Comment:
warning: integer literal has suffix 'u', which is not uppercase
[readability-uppercase-literal-suffix]
```suggestion
return std::tuple {probe_idx, 0U, 0};
```
##########
be/src/vec/common/hash_table/hash_map.h:
##########
@@ -193,10 +197,214 @@
bool has_null_key_data() const { return false; }
};
+template <typename Key, typename Cell, typename Hash = DefaultHash<Key>,
+ typename Grower = HashTableGrower<>, typename Allocator =
HashTableAllocator>
+class JoinHashMapTable : public HashMapTable<Key, Cell, Hash, Grower,
Allocator> {
+public:
+ using Self = JoinHashMapTable;
+ using Base = HashMapTable<Key, Cell, Hash, Grower, Allocator>;
+
+ using key_type = Key;
+ using value_type = typename Cell::value_type;
+ using mapped_type = typename Cell::Mapped;
+
+ using LookupResult = typename Base::LookupResult;
+
+ using HashMapTable<Key, Cell, Hash, Grower, Allocator>::HashMapTable;
+
+ static uint32_t calc_bucket_size(size_t num_elem) {
+ size_t expect_bucket_size = static_cast<size_t>(num_elem) + (num_elem
- 1) / 7;
+ return phmap::priv::NormalizeCapacity(expect_bucket_size) + 1;
+ }
+
+ template <int JoinOpType>
+ void prepare_build(size_t num_elem, int batch_size) {
+ max_batch_size = batch_size;
+ bucket_size = calc_bucket_size(num_elem + 1);
+ first.resize(bucket_size, 0);
+ next.resize(num_elem);
+
+ if constexpr (JoinOpType == doris::TJoinOp::FULL_OUTER_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_OUTER_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_ANTI_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_SEMI_JOIN) {
+ visited.resize(num_elem, 0);
+ }
+ }
+
+ uint32_t get_bucket_size() const { return bucket_size; }
+
+ size_t size() const { return next.size(); }
+
+ void build(const Key* __restrict keys, const uint32_t* __restrict
bucket_nums,
+ size_t num_elem) {
+ build_keys = keys;
+ for (size_t i = 1; i < num_elem; i++) {
+ uint32_t bucket_num = bucket_nums[i];
+ next[i] = first[bucket_num];
+ first[bucket_num] = i;
+ }
+ }
+
+ template <int JoinOpType>
+ auto find_batch(const Key* __restrict keys, const uint32_t* __restrict
bucket_nums,
+ int probe_idx, uint32_t build_idx, int probe_rows,
+ uint32_t* __restrict probe_idxs, uint32_t* __restrict
build_idxs) {
+ if constexpr (JoinOpType == doris::TJoinOp::INNER_JOIN ||
+ JoinOpType == doris::TJoinOp::FULL_OUTER_JOIN ||
+ JoinOpType == doris::TJoinOp::LEFT_OUTER_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_OUTER_JOIN) {
+ return _find_batch_inner_outer_join<JoinOpType>(keys, bucket_nums,
probe_idx, build_idx,
+ probe_rows,
probe_idxs, build_idxs);
+ }
+ if constexpr (JoinOpType == doris::TJoinOp::LEFT_ANTI_JOIN ||
+ JoinOpType == doris::TJoinOp::LEFT_SEMI_JOIN) {
+ return _find_batch_left_semi_anti<JoinOpType>(keys, bucket_nums,
probe_idx, probe_rows,
+ probe_idxs);
+ }
+ if constexpr (JoinOpType == doris::TJoinOp::RIGHT_ANTI_JOIN ||
+ JoinOpType == doris::TJoinOp::RIGHT_SEMI_JOIN) {
+ return _find_batch_right_semi_anti(keys, bucket_nums, probe_idx,
probe_rows);
+ }
+ return std::tuple {0, 0u, 0};
+ }
+
+ template <int JoinOpType>
+ bool iterate_map(std::vector<uint32_t>& build_idxs) const {
+ const auto batch_size = max_batch_size;
+ const auto elem_num = visited.size();
+ int count = 0;
+ build_idxs.resize(batch_size);
+
+ while (count < batch_size && iter_idx < elem_num) {
+ const auto matched = visited[iter_idx];
+ build_idxs[count] = iter_idx;
+ if constexpr (JoinOpType == doris::TJoinOp::RIGHT_ANTI_JOIN) {
+ count += !matched;
+ } else {
+ count += matched;
+ }
+ iter_idx++;
+ }
+
+ build_idxs.resize(count);
+ return iter_idx >= elem_num;
+ }
+
+private:
+ auto _find_batch_right_semi_anti(const Key* __restrict keys,
+ const uint32_t* __restrict bucket_nums,
int probe_idx,
+ int probe_rows) {
+ while (probe_idx < probe_rows) {
+ auto build_idx = first[bucket_nums[probe_idx]];
+
+ while (build_idx) {
+ if (keys[probe_idx] == build_keys[build_idx]) {
+ visited[build_idx] = 1;
+ }
+ build_idx = next[build_idx];
+ }
+ probe_idx++;
+ }
+ return std::tuple {probe_idx, 0u, 0};
+ }
+
+ template <int JoinOpType>
+ auto _find_batch_left_semi_anti(const Key* __restrict keys,
+ const uint32_t* __restrict bucket_nums,
int probe_idx,
+ int probe_rows, uint32_t* __restrict
probe_idxs) {
+ auto matched_cnt = 0;
+ const auto batch_size = max_batch_size;
+
+ while (probe_idx < probe_rows && matched_cnt < batch_size) {
+ uint32_t bucket_num = bucket_nums[probe_idx];
+ auto build_idx = first[bucket_num];
+
+ while (build_idx) {
+ if (keys[probe_idx] == build_keys[build_idx]) {
+ break;
+ }
+ build_idx = next[build_idx];
+ }
+ bool matched =
+ JoinOpType == doris::TJoinOp::LEFT_SEMI_JOIN ? build_idx
!= 0 : build_idx == 0;
+ matched_cnt += matched;
+ probe_idxs[matched_cnt - matched] = probe_idx++;
+ }
+ return std::tuple {probe_idx, 0u, matched_cnt};
Review Comment:
warning: integer literal has suffix 'u', which is not uppercase
[readability-uppercase-literal-suffix]
```suggestion
return std::tuple {probe_idx, 0U, matched_cnt};
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org
For additional commands, e-mail: commits-h...@doris.apache.org
