Coverage Report

Created: 2026-06-05 06:00

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
29
template <typename ChannelIds>
30
300k
Status Crc32HashPartitioner<ChannelIds>::do_partitioning(RuntimeState* state, Block* block) const {
31
300k
    size_t rows = block->rows();
32
33
300k
    if (rows > 0) {
34
47.4k
        auto column_to_keep = block->columns();
35
36
47.4k
        int result_size = cast_set<int>(_partition_expr_ctxs.size());
37
47.4k
        std::vector<int> result(result_size);
38
39
47.4k
        _initialize_hash_vals(rows);
40
47.4k
        auto* __restrict hashes = _hash_vals.data();
41
47.4k
        RETURN_IF_ERROR(_get_partition_column_result(block, result));
42
117k
        for (int j = 0; j < result_size; ++j) {
43
70.2k
            const auto& [col, is_const] = unpack_if_const(block->get_by_position(result[j]).column);
44
70.2k
            if (is_const) {
45
3
                continue;
46
3
            }
47
70.2k
            _do_hash(col, hashes, j);
48
70.2k
        }
49
50
42.8M
        for (size_t i = 0; i < rows; i++) {
51
42.8M
            hashes[i] = ChannelIds()(hashes[i], _partition_count);
52
42.8M
        }
53
54
47.4k
        Block::erase_useless_column(block, column_to_keep);
55
47.4k
    }
56
300k
    return Status::OK();
57
300k
}
_ZNK5doris20Crc32HashPartitionerINS_17ShuffleChannelIdsEE15do_partitioningEPNS_12RuntimeStateEPNS_5BlockE
Line
Count
Source
30
9.12k
Status Crc32HashPartitioner<ChannelIds>::do_partitioning(RuntimeState* state, Block* block) const {
31
9.12k
    size_t rows = block->rows();
32
33
9.12k
    if (rows > 0) {
34
4.48k
        auto column_to_keep = block->columns();
35
36
4.48k
        int result_size = cast_set<int>(_partition_expr_ctxs.size());
37
4.48k
        std::vector<int> result(result_size);
38
39
4.48k
        _initialize_hash_vals(rows);
40
4.48k
        auto* __restrict hashes = _hash_vals.data();
41
4.48k
        RETURN_IF_ERROR(_get_partition_column_result(block, result));
42
8.97k
        for (int j = 0; j < result_size; ++j) {
43
4.49k
            const auto& [col, is_const] = unpack_if_const(block->get_by_position(result[j]).column);
44
4.49k
            if (is_const) {
45
0
                continue;
46
0
            }
47
4.49k
            _do_hash(col, hashes, j);
48
4.49k
        }
49
50
7.91M
        for (size_t i = 0; i < rows; i++) {
51
7.90M
            hashes[i] = ChannelIds()(hashes[i], _partition_count);
52
7.90M
        }
53
54
4.48k
        Block::erase_useless_column(block, column_to_keep);
55
4.48k
    }
56
9.12k
    return Status::OK();
57
9.12k
}
_ZNK5doris20Crc32HashPartitionerINS_24SpillPartitionChannelIdsEE15do_partitioningEPNS_12RuntimeStateEPNS_5BlockE
Line
Count
Source
30
8
Status Crc32HashPartitioner<ChannelIds>::do_partitioning(RuntimeState* state, Block* block) const {
31
8
    size_t rows = block->rows();
32
33
8
    if (rows > 0) {
34
8
        auto column_to_keep = block->columns();
35
36
8
        int result_size = cast_set<int>(_partition_expr_ctxs.size());
37
8
        std::vector<int> result(result_size);
38
39
8
        _initialize_hash_vals(rows);
40
8
        auto* __restrict hashes = _hash_vals.data();
41
8
        RETURN_IF_ERROR(_get_partition_column_result(block, result));
42
16
        for (int j = 0; j < result_size; ++j) {
43
8
            const auto& [col, is_const] = unpack_if_const(block->get_by_position(result[j]).column);
44
8
            if (is_const) {
45
0
                continue;
46
0
            }
47
8
            _do_hash(col, hashes, j);
48
8
        }
49
50
3.14M
        for (size_t i = 0; i < rows; i++) {
51
3.14M
            hashes[i] = ChannelIds()(hashes[i], _partition_count);
52
3.14M
        }
53
54
8
        Block::erase_useless_column(block, column_to_keep);
55
8
    }
56
8
    return Status::OK();
57
8
}
_ZNK5doris20Crc32HashPartitionerINS_26SpillRePartitionChannelIdsEE15do_partitioningEPNS_12RuntimeStateEPNS_5BlockE
Line
Count
Source
30
8
Status Crc32HashPartitioner<ChannelIds>::do_partitioning(RuntimeState* state, Block* block) const {
31
8
    size_t rows = block->rows();
32
33
8
    if (rows > 0) {
34
8
        auto column_to_keep = block->columns();
35
36
8
        int result_size = cast_set<int>(_partition_expr_ctxs.size());
37
8
        std::vector<int> result(result_size);
38
39
8
        _initialize_hash_vals(rows);
40
8
        auto* __restrict hashes = _hash_vals.data();
41
8
        RETURN_IF_ERROR(_get_partition_column_result(block, result));
42
16
        for (int j = 0; j < result_size; ++j) {
43
8
            const auto& [col, is_const] = unpack_if_const(block->get_by_position(result[j]).column);
44
8
            if (is_const) {
45
0
                continue;
46
0
            }
47
8
            _do_hash(col, hashes, j);
48
8
        }
49
50
28
        for (size_t i = 0; i < rows; i++) {
51
20
            hashes[i] = ChannelIds()(hashes[i], _partition_count);
52
20
        }
53
54
8
        Block::erase_useless_column(block, column_to_keep);
55
8
    }
56
8
    return Status::OK();
57
8
}
_ZNK5doris20Crc32HashPartitionerINS_15ShiftChannelIdsEE15do_partitioningEPNS_12RuntimeStateEPNS_5BlockE
Line
Count
Source
30
291k
Status Crc32HashPartitioner<ChannelIds>::do_partitioning(RuntimeState* state, Block* block) const {
31
291k
    size_t rows = block->rows();
32
33
291k
    if (rows > 0) {
34
42.9k
        auto column_to_keep = block->columns();
35
36
42.9k
        int result_size = cast_set<int>(_partition_expr_ctxs.size());
37
42.9k
        std::vector<int> result(result_size);
38
39
42.9k
        _initialize_hash_vals(rows);
40
42.9k
        auto* __restrict hashes = _hash_vals.data();
41
42.9k
        RETURN_IF_ERROR(_get_partition_column_result(block, result));
42
108k
        for (int j = 0; j < result_size; ++j) {
43
65.7k
            const auto& [col, is_const] = unpack_if_const(block->get_by_position(result[j]).column);
44
65.7k
            if (is_const) {
45
3
                continue;
46
3
            }
47
65.7k
            _do_hash(col, hashes, j);
48
65.7k
        }
49
50
31.8M
        for (size_t i = 0; i < rows; i++) {
51
31.7M
            hashes[i] = ChannelIds()(hashes[i], _partition_count);
52
31.7M
        }
53
54
42.9k
        Block::erase_useless_column(block, column_to_keep);
55
42.9k
    }
56
291k
    return Status::OK();
57
291k
}
58
59
template <typename ChannelIds>
60
void Crc32HashPartitioner<ChannelIds>::_do_hash(const ColumnPtr& column,
61
4.50k
                                                HashValType* __restrict result, int idx) const {
62
4.50k
    column->update_crcs_with_value(
63
4.50k
            result, _partition_expr_ctxs[idx]->root()->data_type()->get_primitive_type(),
64
4.50k
            cast_set<HashValType>(column->size()));
65
4.50k
}
_ZNK5doris20Crc32HashPartitionerINS_17ShuffleChannelIdsEE8_do_hashERKNS_3COWINS_7IColumnEE13immutable_ptrIS4_EEPji
Line
Count
Source
61
4.48k
                                                HashValType* __restrict result, int idx) const {
62
4.48k
    column->update_crcs_with_value(
63
4.48k
            result, _partition_expr_ctxs[idx]->root()->data_type()->get_primitive_type(),
64
4.48k
            cast_set<HashValType>(column->size()));
65
4.48k
}
_ZNK5doris20Crc32HashPartitionerINS_24SpillPartitionChannelIdsEE8_do_hashERKNS_3COWINS_7IColumnEE13immutable_ptrIS4_EEPji
Line
Count
Source
61
8
                                                HashValType* __restrict result, int idx) const {
62
8
    column->update_crcs_with_value(
63
8
            result, _partition_expr_ctxs[idx]->root()->data_type()->get_primitive_type(),
64
8
            cast_set<HashValType>(column->size()));
65
8
}
_ZNK5doris20Crc32HashPartitionerINS_26SpillRePartitionChannelIdsEE8_do_hashERKNS_3COWINS_7IColumnEE13immutable_ptrIS4_EEPji
Line
Count
Source
61
8
                                                HashValType* __restrict result, int idx) const {
62
8
    column->update_crcs_with_value(
63
8
            result, _partition_expr_ctxs[idx]->root()->data_type()->get_primitive_type(),
64
8
            cast_set<HashValType>(column->size()));
65
8
}
Unexecuted instantiation: _ZNK5doris20Crc32HashPartitionerINS_15ShiftChannelIdsEE8_do_hashERKNS_3COWINS_7IColumnEE13immutable_ptrIS4_EEPji
66
67
template <typename ChannelIds>
68
Status Crc32HashPartitioner<ChannelIds>::clone(RuntimeState* state,
69
2.78k
                                               std::unique_ptr<PartitionerBase>& partitioner) {
70
2.78k
    auto* new_partitioner = new Crc32HashPartitioner<ChannelIds>(_partition_count);
71
2.78k
    partitioner.reset(new_partitioner);
72
2.78k
    return _clone_expr_ctxs(state, new_partitioner->_partition_expr_ctxs);
73
2.78k
}
_ZN5doris20Crc32HashPartitionerINS_17ShuffleChannelIdsEE5cloneEPNS_12RuntimeStateERSt10unique_ptrINS_15PartitionerBaseESt14default_deleteIS6_EE
Line
Count
Source
69
2.77k
                                               std::unique_ptr<PartitionerBase>& partitioner) {
70
2.77k
    auto* new_partitioner = new Crc32HashPartitioner<ChannelIds>(_partition_count);
71
2.77k
    partitioner.reset(new_partitioner);
72
2.77k
    return _clone_expr_ctxs(state, new_partitioner->_partition_expr_ctxs);
73
2.77k
}
_ZN5doris20Crc32HashPartitionerINS_24SpillPartitionChannelIdsEE5cloneEPNS_12RuntimeStateERSt10unique_ptrINS_15PartitionerBaseESt14default_deleteIS6_EE
Line
Count
Source
69
5
                                               std::unique_ptr<PartitionerBase>& partitioner) {
70
5
    auto* new_partitioner = new Crc32HashPartitioner<ChannelIds>(_partition_count);
71
5
    partitioner.reset(new_partitioner);
72
5
    return _clone_expr_ctxs(state, new_partitioner->_partition_expr_ctxs);
73
5
}
_ZN5doris20Crc32HashPartitionerINS_26SpillRePartitionChannelIdsEE5cloneEPNS_12RuntimeStateERSt10unique_ptrINS_15PartitionerBaseESt14default_deleteIS6_EE
Line
Count
Source
69
4
                                               std::unique_ptr<PartitionerBase>& partitioner) {
70
4
    auto* new_partitioner = new Crc32HashPartitioner<ChannelIds>(_partition_count);
71
4
    partitioner.reset(new_partitioner);
72
4
    return _clone_expr_ctxs(state, new_partitioner->_partition_expr_ctxs);
73
4
}
Unexecuted instantiation: _ZN5doris20Crc32HashPartitionerINS_15ShiftChannelIdsEE5cloneEPNS_12RuntimeStateERSt10unique_ptrINS_15PartitionerBaseESt14default_deleteIS6_EE
74
75
void Crc32CHashPartitioner::_do_hash(const ColumnPtr& column, HashValType* __restrict result,
76
65.7k
                                     int idx) const {
77
65.7k
    column->update_crc32c_batch(result, nullptr);
78
65.7k
}
79
80
Status Crc32CHashPartitioner::clone(RuntimeState* state,
81
166k
                                    std::unique_ptr<PartitionerBase>& partitioner) {
82
166k
    auto* new_partitioner = new Crc32CHashPartitioner(_partition_count);
83
166k
    partitioner.reset(new_partitioner);
84
166k
    return _clone_expr_ctxs(state, new_partitioner->_partition_expr_ctxs);
85
166k
}
86
87
template class Crc32HashPartitioner<ShuffleChannelIds>;
88
template class Crc32HashPartitioner<SpillPartitionChannelIds>;
89
template class Crc32HashPartitioner<SpillRePartitionChannelIds>;
90
91
} // namespace doris