Coverage Report

Created: 2026-10-29 14:40

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
be/src/exprs/aggregate/aggregate_function_count.h
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
// This file is copied from
18
// https://github.com/ClickHouse/ClickHouse/blob/master/src/AggregateFunctions/AggregateFunctionCount.h
19
// and modified by Doris
20
21
#pragma once
22
23
#include <stddef.h>
24
25
#include <algorithm>
26
#include <boost/iterator/iterator_facade.hpp>
27
#include <memory>
28
#include <vector>
29
30
#include "core/assert_cast.h"
31
#include "core/column/column.h"
32
#include "core/column/column_fixed_length_object.h"
33
#include "core/column/column_nullable.h"
34
#include "core/column/column_vector.h"
35
#include "core/data_type/data_type.h"
36
#include "core/data_type/data_type_fixed_length_object.h"
37
#include "core/data_type/data_type_number.h"
38
#include "core/types.h"
39
#include "exprs/aggregate/aggregate_function.h"
40
#include "util/simd/bits.h"
41
42
namespace doris {
43
class Arena;
44
class BufferReadable;
45
class BufferWritable;
46
47
struct AggregateFunctionCountData {
48
    UInt64 count = 0;
49
};
50
51
/// Simply count number of calls.
52
class AggregateFunctionCount final
53
        : public IAggregateFunctionDataHelper<AggregateFunctionCountData, AggregateFunctionCount>,
54
          VarargsExpression,
55
          NotNullableAggregateFunction {
56
public:
57
    AggregateFunctionCount(const DataTypes& argument_types_)
58
30.5k
            : IAggregateFunctionDataHelper(argument_types_) {}
59
60
5.31k
    bool is_simple_count() const override { return true; }
61
388
    String get_name() const override { return "count"; }
62
63
115k
    DataTypePtr get_return_type() const override { return std::make_shared<DataTypeInt64>(); }
64
65
1
    bool is_trivial() const override { return true; }
66
67
545
    WindowSpillStrategy window_spill_strategy() const override {
68
545
        return WindowSpillStrategy::PARTITION_REDUCE;
69
545
    }
70
71
13.4M
    void add(AggregateDataPtr __restrict place, const IColumn**, ssize_t, Arena&) const override {
72
13.4M
        ++data(place).count;
73
13.4M
    }
74
75
    void add_batch_single_place(size_t batch_size, AggregateDataPtr place, const IColumn**,
76
58.6k
                                Arena&) const override {
77
58.6k
        data(place).count += batch_size;
78
58.6k
    }
79
80
309
    void reset(AggregateDataPtr place) const override {
81
309
        AggregateFunctionCount::data(place).count = 0;
82
309
    }
83
84
    void merge(AggregateDataPtr __restrict place, ConstAggregateDataPtr rhs,
85
111k
               Arena&) const override {
86
111k
        data(place).count += data(rhs).count;
87
111k
    }
88
89
68
    void serialize(ConstAggregateDataPtr __restrict place, BufferWritable& buf) const override {
90
68
        buf.write_var_uint(data(place).count);
91
68
    }
92
93
    void deserialize(AggregateDataPtr __restrict place, BufferReadable& buf,
94
68
                     Arena&) const override {
95
68
        buf.read_var_uint(data(place).count);
96
68
    }
97
98
986k
    void insert_result_into(ConstAggregateDataPtr __restrict place, IColumn& to) const override {
99
986k
        assert_cast<ColumnInt64&, TypeCheckOnRelease::DISABLE>(to).get_data().push_back(
100
986k
                data(place).count);
101
986k
    }
102
103
    void serialize_to_column(const std::vector<AggregateDataPtr>& places, size_t offset,
104
4.52k
                             MutableColumnPtr& dst, const size_t num_rows) const override {
105
4.52k
        auto& col = assert_cast<ColumnFixedLengthObject&>(*dst);
106
18.4E
        DCHECK(col.item_size() == sizeof(Data))
107
18.4E
                << "size is not equal: " << col.item_size() << " " << sizeof(Data);
108
4.52k
        col.resize(num_rows);
109
4.52k
        auto* data = col.get_data().data();
110
216k
        for (size_t i = 0; i != num_rows; ++i) {
111
212k
            *reinterpret_cast<Data*>(&data[sizeof(Data) * i]) =
112
212k
                    *reinterpret_cast<Data*>(places[i] + offset);
113
212k
        }
114
4.52k
    }
115
116
    void streaming_agg_serialize_to_column(const IColumn** columns, MutableColumnPtr& dst,
117
61
                                           const size_t num_rows, Arena&) const override {
118
61
        auto& dst_col = assert_cast<ColumnFixedLengthObject&>(*dst);
119
61
        DCHECK(dst_col.item_size() == sizeof(Data))
120
1
                << "size is not equal: " << dst_col.item_size() << " " << sizeof(Data);
121
61
        dst_col.resize(num_rows);
122
61
        auto* data = dst_col.get_data().data();
123
336
        for (size_t i = 0; i != num_rows; ++i) {
124
275
            auto& state = *reinterpret_cast<Data*>(&data[sizeof(Data) * i]);
125
275
            state.count = 1;
126
275
        }
127
61
    }
128
129
    void deserialize_from_column_row(AggregateDataPtr __restrict place, const IColumn& column,
130
12
                                     size_t row, Arena&) const override {
131
12
        DCHECK_LT(row, column.size());
132
12
        const auto& col = assert_cast<const ColumnFixedLengthObject&>(column);
133
12
        const auto* data = reinterpret_cast<const Data*>(col.get_data().data());
134
12
        AggregateFunctionCount::data(place) = data[row];
135
12
    }
136
137
    void deserialize_and_merge_from_column_range(AggregateDataPtr __restrict place,
138
                                                 const IColumn& column, size_t begin, size_t end,
139
6.16k
                                                 Arena&) const override {
140
6.16k
        DCHECK(end <= column.size() && begin <= end)
141
0
                << ", begin:" << begin << ", end:" << end << ", column.size():" << column.size();
142
6.16k
        auto& col = assert_cast<const ColumnFixedLengthObject&>(column);
143
6.16k
        auto* data = reinterpret_cast<const Data*>(col.get_data().data());
144
115k
        for (size_t i = begin; i <= end; ++i) {
145
108k
            doris::AggregateFunctionCount::data(place).count += data[i].count;
146
108k
        }
147
6.16k
    }
148
149
131
    bool needs_deserialize_and_merge_scratch() const override { return false; }
150
151
    void deserialize_and_merge_vec(const AggregateDataPtr* places, size_t offset,
152
                                   AggregateDataPtr rhs, const IColumn* column, Arena& arena,
153
3.39k
                                   const size_t num_rows) const override {
154
3.39k
        const auto& col = assert_cast<const ColumnFixedLengthObject&>(*column);
155
3.39k
        const auto* data = col.get_data().data();
156
3.39k
        this->merge_vec(places, offset, AggregateDataPtr(data), arena, num_rows);
157
3.39k
    }
158
159
    void deserialize_and_merge_vec_selected(const AggregateDataPtr* places, size_t offset,
160
                                            AggregateDataPtr rhs, const IColumn* column,
161
21
                                            Arena& arena, const size_t num_rows) const override {
162
21
        const auto& col = assert_cast<const ColumnFixedLengthObject&>(*column);
163
21
        const auto* data = col.get_data().data();
164
21
        this->merge_vec_selected(places, offset, AggregateDataPtr(data), arena, num_rows);
165
21
    }
166
167
    void serialize_without_key_to_column(ConstAggregateDataPtr __restrict place,
168
5.85k
                                         IColumn& to) const override {
169
5.85k
        auto& col = assert_cast<ColumnFixedLengthObject&>(to);
170
5.85k
        DCHECK(col.item_size() == sizeof(Data))
171
10
                << "size is not equal: " << col.item_size() << " " << sizeof(Data);
172
5.85k
        size_t old_size = col.size();
173
5.85k
        col.resize(old_size + 1);
174
5.85k
        (reinterpret_cast<Data*>(col.get_data().data()) + old_size)->count =
175
5.85k
                AggregateFunctionCount::data(place).count;
176
5.85k
    }
177
178
11.3k
    MutableColumnPtr create_serialize_column() const override {
179
11.3k
        return ColumnFixedLengthObject::create(sizeof(Data));
180
11.3k
    }
181
182
11.5k
    DataTypePtr get_serialized_type() const override {
183
11.5k
        return std::make_shared<DataTypeFixedLengthObject>();
184
11.5k
    }
185
186
    void add_range_single_place(int64_t partition_start, int64_t partition_end, int64_t frame_start,
187
                                int64_t frame_end, AggregateDataPtr place, const IColumn** columns,
188
                                Arena& arena, UInt8* use_null_result,
189
1.23k
                                UInt8* could_use_previous_result) const override {
190
1.23k
        frame_start = std::max<int64_t>(frame_start, partition_start);
191
1.23k
        frame_end = std::min<int64_t>(frame_end, partition_end);
192
1.23k
        if (frame_start >= frame_end) {
193
19
            if (!*could_use_previous_result) {
194
2
                *use_null_result = true;
195
2
            }
196
1.21k
        } else {
197
1.21k
            AggregateFunctionCount::data(place).count += frame_end - frame_start;
198
1.21k
            *use_null_result = false;
199
1.21k
            *could_use_previous_result = true;
200
1.21k
        }
201
1.23k
    }
202
};
203
204
// Used for unary count(nullable_expr). SQL count(expr) counts non-NULL values.
205
class AggregateFunctionCountNotNullUnary final
206
        : public IAggregateFunctionDataHelper<AggregateFunctionCountData,
207
                                              AggregateFunctionCountNotNullUnary> {
208
public:
209
    AggregateFunctionCountNotNullUnary(const DataTypes& argument_types_)
210
14.3k
            : IAggregateFunctionDataHelper(argument_types_) {}
211
212
380
    String get_name() const override { return "count"; }
213
214
108k
    DataTypePtr get_return_type() const override { return std::make_shared<DataTypeInt64>(); }
215
216
7
    bool is_trivial() const override { return true; }
217
218
555
    WindowSpillStrategy window_spill_strategy() const override {
219
555
        return WindowSpillStrategy::PARTITION_REDUCE;
220
555
    }
221
222
    void add(AggregateDataPtr __restrict place, const IColumn** columns, ssize_t row_num,
223
2.12M
             Arena&) const override {
224
2.12M
        data(place).count +=
225
2.12M
                !assert_cast<const ColumnNullable&, TypeCheckOnRelease::DISABLE>(*columns[0])
226
2.12M
                         .is_null_at(row_num);
227
2.12M
    }
228
229
    void add_batch_single_place(size_t batch_size, AggregateDataPtr place, const IColumn** columns,
230
16.2k
                                Arena&) const override {
231
16.2k
        const auto& nullable_column =
232
16.2k
                assert_cast<const ColumnNullable&, TypeCheckOnRelease::DISABLE>(*columns[0]);
233
16.2k
        const auto& null_map = nullable_column.get_null_map_data();
234
16.2k
        DCHECK_LE(batch_size, null_map.size());
235
16.2k
        if (!nullable_column.has_null(0, batch_size)) {
236
12.9k
            data(place).count += batch_size;
237
12.9k
            return;
238
12.9k
        }
239
3.33k
        data(place).count +=
240
3.33k
                simd::count_zero_num(reinterpret_cast<const int8_t*>(null_map.data()), batch_size);
241
3.33k
    }
242
243
60.1k
    void reset(AggregateDataPtr place) const override { data(place).count = 0; }
244
245
    void merge(AggregateDataPtr __restrict place, ConstAggregateDataPtr rhs,
246
88.8k
               Arena&) const override {
247
88.8k
        data(place).count += data(rhs).count;
248
88.8k
    }
249
250
257
    void serialize(ConstAggregateDataPtr __restrict place, BufferWritable& buf) const override {
251
257
        buf.write_var_uint(data(place).count);
252
257
    }
253
254
    void deserialize(AggregateDataPtr __restrict place, BufferReadable& buf,
255
257
                     Arena&) const override {
256
257
        buf.read_var_uint(data(place).count);
257
257
    }
258
259
178k
    void insert_result_into(ConstAggregateDataPtr __restrict place, IColumn& to) const override {
260
178k
        if (is_column_nullable(to)) {
261
1
            auto& null_column = assert_cast<ColumnNullable&, TypeCheckOnRelease::DISABLE>(to);
262
1
            null_column.get_null_map_data().push_back(0);
263
1
            assert_cast<ColumnInt64&, TypeCheckOnRelease::DISABLE>(null_column.get_nested_column())
264
1
                    .get_data()
265
1
                    .push_back(data(place).count);
266
178k
        } else {
267
178k
            assert_cast<ColumnInt64&, TypeCheckOnRelease::DISABLE>(to).get_data().push_back(
268
178k
                    data(place).count);
269
178k
        }
270
178k
    }
271
272
74.1k
    void check_result_column_type(const IColumn& to) const override {
273
74.1k
        if (const auto* null_column = check_and_get_column<ColumnNullable>(to)) {
274
1
            IAggregateFunction::check_result_column_type(null_column->get_nested_column());
275
1
            return;
276
1
        }
277
74.1k
        IAggregateFunction::check_result_column_type(to);
278
74.1k
    }
279
280
    void serialize_to_column(const std::vector<AggregateDataPtr>& places, size_t offset,
281
3.51k
                             MutableColumnPtr& dst, const size_t num_rows) const override {
282
3.51k
        auto& col = assert_cast<ColumnFixedLengthObject&>(*dst);
283
18.4E
        DCHECK(col.item_size() == sizeof(Data))
284
18.4E
                << "size is not equal: " << col.item_size() << " " << sizeof(Data);
285
3.51k
        col.resize(num_rows);
286
3.51k
        auto* data = col.get_data().data();
287
166k
        for (size_t i = 0; i != num_rows; ++i) {
288
162k
            *reinterpret_cast<Data*>(&data[sizeof(Data) * i]) =
289
162k
                    *reinterpret_cast<Data*>(places[i] + offset);
290
162k
        }
291
3.51k
    }
292
293
    void streaming_agg_serialize_to_column(const IColumn** columns, MutableColumnPtr& dst,
294
73
                                           const size_t num_rows, Arena&) const override {
295
73
        auto& col = assert_cast<ColumnFixedLengthObject&>(*dst);
296
73
        DCHECK(col.item_size() == sizeof(Data))
297
0
                << "size is not equal: " << col.item_size() << " " << sizeof(Data);
298
73
        col.resize(num_rows);
299
73
        auto& data = col.get_data();
300
73
        const ColumnNullable& input_col = assert_cast<const ColumnNullable&>(*columns[0]);
301
223
        for (size_t i = 0; i < num_rows; i++) {
302
150
            auto& state = *reinterpret_cast<Data*>(&data[sizeof(Data) * i]);
303
150
            state.count = !input_col.is_null_at(i);
304
150
        }
305
73
    }
306
307
    void deserialize_from_column_row(AggregateDataPtr __restrict place, const IColumn& column,
308
38
                                     size_t row, Arena&) const override {
309
38
        DCHECK_LT(row, column.size());
310
38
        const auto& col = assert_cast<const ColumnFixedLengthObject&>(column);
311
38
        const auto* data = reinterpret_cast<const Data*>(col.get_data().data());
312
38
        AggregateFunctionCountNotNullUnary::data(place) = data[row];
313
38
    }
314
315
    void deserialize_and_merge_from_column_range(AggregateDataPtr __restrict place,
316
                                                 const IColumn& column, size_t begin, size_t end,
317
5.52k
                                                 Arena&) const override {
318
5.52k
        DCHECK(end <= column.size() && begin <= end)
319
0
                << ", begin:" << begin << ", end:" << end << ", column.size():" << column.size();
320
5.52k
        auto& col = assert_cast<const ColumnFixedLengthObject&>(column);
321
5.52k
        auto* data = reinterpret_cast<const Data*>(col.get_data().data());
322
85.8k
        for (size_t i = begin; i <= end; ++i) {
323
80.2k
            doris::AggregateFunctionCountNotNullUnary::data(place).count += data[i].count;
324
80.2k
        }
325
5.52k
    }
326
327
105
    bool needs_deserialize_and_merge_scratch() const override { return false; }
328
329
    void deserialize_and_merge_vec(const AggregateDataPtr* places, size_t offset,
330
                                   AggregateDataPtr rhs, const IColumn* column, Arena& arena,
331
2.34k
                                   const size_t num_rows) const override {
332
2.34k
        const auto& col = assert_cast<const ColumnFixedLengthObject&>(*column);
333
2.34k
        const auto* data = col.get_data().data();
334
2.34k
        this->merge_vec(places, offset, AggregateDataPtr(data), arena, num_rows);
335
2.34k
    }
336
337
    void deserialize_and_merge_vec_selected(const AggregateDataPtr* places, size_t offset,
338
                                            AggregateDataPtr rhs, const IColumn* column,
339
22
                                            Arena& arena, const size_t num_rows) const override {
340
22
        const auto& col = assert_cast<const ColumnFixedLengthObject&>(*column);
341
22
        const auto* data = col.get_data().data();
342
22
        this->merge_vec_selected(places, offset, AggregateDataPtr(data), arena, num_rows);
343
22
    }
344
345
    void serialize_without_key_to_column(ConstAggregateDataPtr __restrict place,
346
5.09k
                                         IColumn& to) const override {
347
5.09k
        auto& col = assert_cast<ColumnFixedLengthObject&>(to);
348
18.4E
        DCHECK(col.item_size() == sizeof(Data))
349
18.4E
                << "size is not equal: " << col.item_size() << " " << sizeof(Data);
350
5.09k
        size_t old_size = col.size();
351
5.09k
        col.resize(old_size + 1);
352
5.09k
        (reinterpret_cast<Data*>(col.get_data().data()) + old_size)->count =
353
5.09k
                AggregateFunctionCountNotNullUnary::data(place).count;
354
5.09k
    }
355
356
8.54k
    MutableColumnPtr create_serialize_column() const override {
357
8.54k
        return ColumnFixedLengthObject::create(sizeof(Data));
358
8.54k
    }
359
360
9.08k
    DataTypePtr get_serialized_type() const override {
361
9.08k
        return std::make_shared<DataTypeFixedLengthObject>();
362
9.08k
    }
363
364
    void add_range_single_place(int64_t partition_start, int64_t partition_end, int64_t frame_start,
365
                                int64_t frame_end, AggregateDataPtr place, const IColumn** columns,
366
                                Arena& arena, UInt8* use_null_result,
367
60.5k
                                UInt8* could_use_previous_result) const override {
368
60.5k
        frame_start = std::max<int64_t>(frame_start, partition_start);
369
60.5k
        frame_end = std::min<int64_t>(frame_end, partition_end);
370
60.5k
        if (frame_start >= frame_end) {
371
12
            if (!*could_use_previous_result) {
372
0
                *use_null_result = true;
373
0
            }
374
60.5k
        } else {
375
60.5k
            const auto& nullable_column =
376
60.5k
                    assert_cast<const ColumnNullable&, TypeCheckOnRelease::DISABLE>(*columns[0]);
377
60.5k
            size_t count = 0;
378
60.5k
            if (nullable_column.has_null()) {
379
119k
                for (int64_t i = frame_start; i < frame_end; ++i) {
380
63.8k
                    if (!nullable_column.is_null_at(i)) {
381
61.2k
                        ++count;
382
61.2k
                    }
383
63.8k
                }
384
55.9k
            } else {
385
4.63k
                count = frame_end - frame_start;
386
4.63k
            }
387
60.5k
            *use_null_result = false;
388
60.5k
            *could_use_previous_result = true;
389
60.5k
            AggregateFunctionCountNotNullUnary::data(place).count += count;
390
60.5k
        }
391
60.5k
    }
392
};
393
394
} // namespace doris