Coverage Report

Created: 2026-03-23 10:27

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
be/src/exec/partitioner/partitioner.cpp
Line
Count
Source
1
// Licensed to the Apache Software Foundation (ASF) under one
2
// or more contributor license agreements.  See the NOTICE file
3
// distributed with this work for additional information
4
// regarding copyright ownership.  The ASF licenses this file
5
// to you under the Apache License, Version 2.0 (the
6
// "License"); you may not use this file except in compliance
7
// with the License.  You may obtain a copy of the License at
8
//
9
//   http://www.apache.org/licenses/LICENSE-2.0
10
//
11
// Unless required by applicable law or agreed to in writing,
12
// software distributed under the License is distributed on an
13
// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
14
// KIND, either express or implied.  See the License for the
15
// specific language governing permissions and limitations
16
// under the License.
17
18
#include "exec/partitioner/partitioner.h"
19
20
#include "common/cast_set.h"
21
#include "common/status.h"
22
#include "core/column/column_const.h"
23
#include "exec/exchange/local_exchange_sink_operator.h"
24
#include "exec/exchange/vdata_stream_sender.h"
25
#include "runtime/thread_context.h"
26
27
namespace doris {
28
#include "common/compile_check_begin.h"
29
30
template <typename ChannelIds>
31
286k
Status Crc32HashPartitioner<ChannelIds>::do_partitioning(RuntimeState* state, Block* block) const {
32
286k
    size_t rows = block->rows();
33
34
286k
    if (rows > 0) {
35
46.9k
        auto column_to_keep = block->columns();
36
37
46.9k
        int result_size = cast_set<int>(_partition_expr_ctxs.size());
38
46.9k
        std::vector<int> result(result_size);
39
40
46.9k
        _initialize_hash_vals(rows);
41
46.9k
        auto* __restrict hashes = _hash_vals.data();
42
46.9k
        RETURN_IF_ERROR(_get_partition_column_result(block, result));
43
124k
        for (int j = 0; j < result_size; ++j) {
44
77.5k
            const auto& [col, is_const] = unpack_if_const(block->get_by_position(result[j]).column);
45
77.5k
            if (is_const) {
46
4
                continue;
47
4
            }
48
77.5k
            _do_hash(col, hashes, j);
49
77.5k
        }
50
51
19.3M
        for (size_t i = 0; i < rows; i++) {
52
19.2M
            hashes[i] = ChannelIds()(hashes[i], _partition_count);
53
19.2M
        }
54
55
46.9k
        Block::erase_useless_column(block, column_to_keep);
56
46.9k
    }
57
286k
    return Status::OK();
58
286k
}
_ZNK5doris20Crc32HashPartitionerINS_17ShuffleChannelIdsEE15do_partitioningEPNS_12RuntimeStateEPNS_5BlockE
Line
Count
Source
31
17.9k
Status Crc32HashPartitioner<ChannelIds>::do_partitioning(RuntimeState* state, Block* block) const {
32
17.9k
    size_t rows = block->rows();
33
34
17.9k
    if (rows > 0) {
35
6.01k
        auto column_to_keep = block->columns();
36
37
6.01k
        int result_size = cast_set<int>(_partition_expr_ctxs.size());
38
6.01k
        std::vector<int> result(result_size);
39
40
6.01k
        _initialize_hash_vals(rows);
41
6.01k
        auto* __restrict hashes = _hash_vals.data();
42
6.01k
        RETURN_IF_ERROR(_get_partition_column_result(block, result));
43
12.0k
        for (int j = 0; j < result_size; ++j) {
44
6.02k
            const auto& [col, is_const] = unpack_if_const(block->get_by_position(result[j]).column);
45
6.02k
            if (is_const) {
46
0
                continue;
47
0
            }
48
6.02k
            _do_hash(col, hashes, j);
49
6.02k
        }
50
51
2.60M
        for (size_t i = 0; i < rows; i++) {
52
2.59M
            hashes[i] = ChannelIds()(hashes[i], _partition_count);
53
2.59M
        }
54
55
6.01k
        Block::erase_useless_column(block, column_to_keep);
56
6.01k
    }
57
17.9k
    return Status::OK();
58
17.9k
}
_ZNK5doris20Crc32HashPartitionerINS_24SpillPartitionChannelIdsEE15do_partitioningEPNS_12RuntimeStateEPNS_5BlockE
Line
Count
Source
31
8
Status Crc32HashPartitioner<ChannelIds>::do_partitioning(RuntimeState* state, Block* block) const {
32
8
    size_t rows = block->rows();
33
34
8
    if (rows > 0) {
35
8
        auto column_to_keep = block->columns();
36
37
8
        int result_size = cast_set<int>(_partition_expr_ctxs.size());
38
8
        std::vector<int> result(result_size);
39
40
8
        _initialize_hash_vals(rows);
41
8
        auto* __restrict hashes = _hash_vals.data();
42
8
        RETURN_IF_ERROR(_get_partition_column_result(block, result));
43
16
        for (int j = 0; j < result_size; ++j) {
44
8
            const auto& [col, is_const] = unpack_if_const(block->get_by_position(result[j]).column);
45
8
            if (is_const) {
46
0
                continue;
47
0
            }
48
8
            _do_hash(col, hashes, j);
49
8
        }
50
51
3.14M
        for (size_t i = 0; i < rows; i++) {
52
3.14M
            hashes[i] = ChannelIds()(hashes[i], _partition_count);
53
3.14M
        }
54
55
8
        Block::erase_useless_column(block, column_to_keep);
56
8
    }
57
8
    return Status::OK();
58
8
}
_ZNK5doris20Crc32HashPartitionerINS_26SpillRePartitionChannelIdsEE15do_partitioningEPNS_12RuntimeStateEPNS_5BlockE
Line
Count
Source
31
8
Status Crc32HashPartitioner<ChannelIds>::do_partitioning(RuntimeState* state, Block* block) const {
32
8
    size_t rows = block->rows();
33
34
8
    if (rows > 0) {
35
8
        auto column_to_keep = block->columns();
36
37
8
        int result_size = cast_set<int>(_partition_expr_ctxs.size());
38
8
        std::vector<int> result(result_size);
39
40
8
        _initialize_hash_vals(rows);
41
8
        auto* __restrict hashes = _hash_vals.data();
42
8
        RETURN_IF_ERROR(_get_partition_column_result(block, result));
43
16
        for (int j = 0; j < result_size; ++j) {
44
8
            const auto& [col, is_const] = unpack_if_const(block->get_by_position(result[j]).column);
45
8
            if (is_const) {
46
0
                continue;
47
0
            }
48
8
            _do_hash(col, hashes, j);
49
8
        }
50
51
28
        for (size_t i = 0; i < rows; i++) {
52
20
            hashes[i] = ChannelIds()(hashes[i], _partition_count);
53
20
        }
54
55
8
        Block::erase_useless_column(block, column_to_keep);
56
8
    }
57
8
    return Status::OK();
58
8
}
_ZNK5doris20Crc32HashPartitionerINS_15ShiftChannelIdsEE15do_partitioningEPNS_12RuntimeStateEPNS_5BlockE
Line
Count
Source
31
268k
Status Crc32HashPartitioner<ChannelIds>::do_partitioning(RuntimeState* state, Block* block) const {
32
268k
    size_t rows = block->rows();
33
34
268k
    if (rows > 0) {
35
40.9k
        auto column_to_keep = block->columns();
36
37
40.9k
        int result_size = cast_set<int>(_partition_expr_ctxs.size());
38
40.9k
        std::vector<int> result(result_size);
39
40
40.9k
        _initialize_hash_vals(rows);
41
40.9k
        auto* __restrict hashes = _hash_vals.data();
42
40.9k
        RETURN_IF_ERROR(_get_partition_column_result(block, result));
43
112k
        for (int j = 0; j < result_size; ++j) {
44
71.5k
            const auto& [col, is_const] = unpack_if_const(block->get_by_position(result[j]).column);
45
71.5k
            if (is_const) {
46
4
                continue;
47
4
            }
48
71.5k
            _do_hash(col, hashes, j);
49
71.5k
        }
50
51
13.5M
        for (size_t i = 0; i < rows; i++) {
52
13.5M
            hashes[i] = ChannelIds()(hashes[i], _partition_count);
53
13.5M
        }
54
55
40.9k
        Block::erase_useless_column(block, column_to_keep);
56
40.9k
    }
57
268k
    return Status::OK();
58
268k
}
59
60
template <typename ChannelIds>
61
void Crc32HashPartitioner<ChannelIds>::_do_hash(const ColumnPtr& column,
62
6.03k
                                                HashValType* __restrict result, int idx) const {
63
6.03k
    column->update_crcs_with_value(
64
6.03k
            result, _partition_expr_ctxs[idx]->root()->data_type()->get_primitive_type(),
65
6.03k
            cast_set<HashValType>(column->size()));
66
6.03k
}
_ZNK5doris20Crc32HashPartitionerINS_17ShuffleChannelIdsEE8_do_hashERKNS_3COWINS_7IColumnEE13immutable_ptrIS4_EEPji
Line
Count
Source
62
6.02k
                                                HashValType* __restrict result, int idx) const {
63
6.02k
    column->update_crcs_with_value(
64
6.02k
            result, _partition_expr_ctxs[idx]->root()->data_type()->get_primitive_type(),
65
6.02k
            cast_set<HashValType>(column->size()));
66
6.02k
}
_ZNK5doris20Crc32HashPartitionerINS_24SpillPartitionChannelIdsEE8_do_hashERKNS_3COWINS_7IColumnEE13immutable_ptrIS4_EEPji
Line
Count
Source
62
8
                                                HashValType* __restrict result, int idx) const {
63
8
    column->update_crcs_with_value(
64
8
            result, _partition_expr_ctxs[idx]->root()->data_type()->get_primitive_type(),
65
8
            cast_set<HashValType>(column->size()));
66
8
}
_ZNK5doris20Crc32HashPartitionerINS_26SpillRePartitionChannelIdsEE8_do_hashERKNS_3COWINS_7IColumnEE13immutable_ptrIS4_EEPji
Line
Count
Source
62
8
                                                HashValType* __restrict result, int idx) const {
63
8
    column->update_crcs_with_value(
64
8
            result, _partition_expr_ctxs[idx]->root()->data_type()->get_primitive_type(),
65
8
            cast_set<HashValType>(column->size()));
66
8
}
Unexecuted instantiation: _ZNK5doris20Crc32HashPartitionerINS_15ShiftChannelIdsEE8_do_hashERKNS_3COWINS_7IColumnEE13immutable_ptrIS4_EEPji
67
68
template <typename ChannelIds>
69
Status Crc32HashPartitioner<ChannelIds>::clone(RuntimeState* state,
70
7.01k
                                               std::unique_ptr<PartitionerBase>& partitioner) {
71
7.01k
    auto* new_partitioner = new Crc32HashPartitioner<ChannelIds>(_partition_count);
72
7.01k
    partitioner.reset(new_partitioner);
73
7.01k
    return _clone_expr_ctxs(state, new_partitioner->_partition_expr_ctxs);
74
7.01k
}
_ZN5doris20Crc32HashPartitionerINS_17ShuffleChannelIdsEE5cloneEPNS_12RuntimeStateERSt10unique_ptrINS_15PartitionerBaseESt14default_deleteIS6_EE
Line
Count
Source
70
7.00k
                                               std::unique_ptr<PartitionerBase>& partitioner) {
71
7.00k
    auto* new_partitioner = new Crc32HashPartitioner<ChannelIds>(_partition_count);
72
7.00k
    partitioner.reset(new_partitioner);
73
7.00k
    return _clone_expr_ctxs(state, new_partitioner->_partition_expr_ctxs);
74
7.00k
}
_ZN5doris20Crc32HashPartitionerINS_24SpillPartitionChannelIdsEE5cloneEPNS_12RuntimeStateERSt10unique_ptrINS_15PartitionerBaseESt14default_deleteIS6_EE
Line
Count
Source
70
5
                                               std::unique_ptr<PartitionerBase>& partitioner) {
71
5
    auto* new_partitioner = new Crc32HashPartitioner<ChannelIds>(_partition_count);
72
5
    partitioner.reset(new_partitioner);
73
5
    return _clone_expr_ctxs(state, new_partitioner->_partition_expr_ctxs);
74
5
}
_ZN5doris20Crc32HashPartitionerINS_26SpillRePartitionChannelIdsEE5cloneEPNS_12RuntimeStateERSt10unique_ptrINS_15PartitionerBaseESt14default_deleteIS6_EE
Line
Count
Source
70
4
                                               std::unique_ptr<PartitionerBase>& partitioner) {
71
4
    auto* new_partitioner = new Crc32HashPartitioner<ChannelIds>(_partition_count);
72
4
    partitioner.reset(new_partitioner);
73
4
    return _clone_expr_ctxs(state, new_partitioner->_partition_expr_ctxs);
74
4
}
Unexecuted instantiation: _ZN5doris20Crc32HashPartitionerINS_15ShiftChannelIdsEE5cloneEPNS_12RuntimeStateERSt10unique_ptrINS_15PartitionerBaseESt14default_deleteIS6_EE
75
76
void Crc32CHashPartitioner::_do_hash(const ColumnPtr& column, HashValType* __restrict result,
77
71.5k
                                     int idx) const {
78
71.5k
    column->update_crc32c_batch(result, nullptr);
79
71.5k
}
80
81
Status Crc32CHashPartitioner::clone(RuntimeState* state,
82
121k
                                    std::unique_ptr<PartitionerBase>& partitioner) {
83
121k
    auto* new_partitioner = new Crc32CHashPartitioner(_partition_count);
84
121k
    partitioner.reset(new_partitioner);
85
121k
    return _clone_expr_ctxs(state, new_partitioner->_partition_expr_ctxs);
86
121k
}
87
88
template class Crc32HashPartitioner<ShuffleChannelIds>;
89
template class Crc32HashPartitioner<SpillPartitionChannelIds>;
90
template class Crc32HashPartitioner<SpillRePartitionChannelIds>;
91
92
} // namespace doris