Coverage Report

Created: 2026-10-09 13:02

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
29.3k
            : IAggregateFunctionDataHelper(argument_types_) {}
59
60
4.72k
    bool is_simple_count() const override { return true; }
61
236
    String get_name() const override { return "count"; }
62
63
109k
    DataTypePtr get_return_type() const override { return std::make_shared<DataTypeInt64>(); }
64
65
0
    bool is_trivial() const override { return true; }
66
67
9.90M
    void add(AggregateDataPtr __restrict place, const IColumn**, ssize_t, Arena&) const override {
68
9.90M
        ++data(place).count;
69
9.90M
    }
70
71
    void add_batch_single_place(size_t batch_size, AggregateDataPtr place, const IColumn**,
72
55.2k
                                Arena&) const override {
73
55.2k
        data(place).count += batch_size;
74
55.2k
    }
75
76
149
    void reset(AggregateDataPtr place) const override {
77
149
        AggregateFunctionCount::data(place).count = 0;
78
149
    }
79
80
    void merge(AggregateDataPtr __restrict place, ConstAggregateDataPtr rhs,
81
107k
               Arena&) const override {
82
107k
        data(place).count += data(rhs).count;
83
107k
    }
84
85
55
    void serialize(ConstAggregateDataPtr __restrict place, BufferWritable& buf) const override {
86
55
        buf.write_var_uint(data(place).count);
87
55
    }
88
89
    void deserialize(AggregateDataPtr __restrict place, BufferReadable& buf,
90
55
                     Arena&) const override {
91
55
        buf.read_var_uint(data(place).count);
92
55
    }
93
94
115k
    void insert_result_into(ConstAggregateDataPtr __restrict place, IColumn& to) const override {
95
115k
        assert_cast<ColumnInt64&, TypeCheckOnRelease::DISABLE>(to).get_data().push_back(
96
115k
                data(place).count);
97
115k
    }
98
99
    void serialize_to_column(const std::vector<AggregateDataPtr>& places, size_t offset,
100
4.02k
                             MutableColumnPtr& dst, const size_t num_rows) const override {
101
4.02k
        auto& col = assert_cast<ColumnFixedLengthObject&>(*dst);
102
4.02k
        DCHECK(col.item_size() == sizeof(Data))
103
9
                << "size is not equal: " << col.item_size() << " " << sizeof(Data);
104
4.02k
        col.resize(num_rows);
105
4.02k
        auto* data = col.get_data().data();
106
215k
        for (size_t i = 0; i != num_rows; ++i) {
107
210k
            *reinterpret_cast<Data*>(&data[sizeof(Data) * i]) =
108
210k
                    *reinterpret_cast<Data*>(places[i] + offset);
109
210k
        }
110
4.02k
    }
111
112
    void streaming_agg_serialize_to_column(const IColumn** columns, MutableColumnPtr& dst,
113
63
                                           const size_t num_rows, Arena&) const override {
114
63
        auto& dst_col = assert_cast<ColumnFixedLengthObject&>(*dst);
115
63
        DCHECK(dst_col.item_size() == sizeof(Data))
116
0
                << "size is not equal: " << dst_col.item_size() << " " << sizeof(Data);
117
63
        dst_col.resize(num_rows);
118
63
        auto* data = dst_col.get_data().data();
119
338
        for (size_t i = 0; i != num_rows; ++i) {
120
275
            auto& state = *reinterpret_cast<Data*>(&data[sizeof(Data) * i]);
121
275
            state.count = 1;
122
275
        }
123
63
    }
124
125
    void deserialize_and_merge_from_column_range(AggregateDataPtr __restrict place,
126
                                                 const IColumn& column, size_t begin, size_t end,
127
6.76k
                                                 Arena&) const override {
128
6.76k
        DCHECK(end <= column.size() && begin <= end)
129
0
                << ", begin:" << begin << ", end:" << end << ", column.size():" << column.size();
130
6.76k
        auto& col = assert_cast<const ColumnFixedLengthObject&>(column);
131
6.76k
        auto* data = reinterpret_cast<const Data*>(col.get_data().data());
132
116k
        for (size_t i = begin; i <= end; ++i) {
133
109k
            doris::AggregateFunctionCount::data(place).count += data[i].count;
134
109k
        }
135
6.76k
    }
136
137
    void deserialize_and_merge_vec(const AggregateDataPtr* places, size_t offset,
138
                                   AggregateDataPtr rhs, const IColumn* column, Arena& arena,
139
3.26k
                                   const size_t num_rows) const override {
140
3.26k
        const auto& col = assert_cast<const ColumnFixedLengthObject&>(*column);
141
3.26k
        const auto* data = col.get_data().data();
142
3.26k
        this->merge_vec(places, offset, AggregateDataPtr(data), arena, num_rows);
143
3.26k
    }
144
145
    void deserialize_and_merge_vec_selected(const AggregateDataPtr* places, size_t offset,
146
                                            AggregateDataPtr rhs, const IColumn* column,
147
1
                                            Arena& arena, const size_t num_rows) const override {
148
1
        const auto& col = assert_cast<const ColumnFixedLengthObject&>(*column);
149
1
        const auto* data = col.get_data().data();
150
1
        this->merge_vec_selected(places, offset, AggregateDataPtr(data), arena, num_rows);
151
1
    }
152
153
    void serialize_without_key_to_column(ConstAggregateDataPtr __restrict place,
154
6.17k
                                         IColumn& to) const override {
155
6.17k
        auto& col = assert_cast<ColumnFixedLengthObject&>(to);
156
6.17k
        DCHECK(col.item_size() == sizeof(Data))
157
0
                << "size is not equal: " << col.item_size() << " " << sizeof(Data);
158
6.17k
        size_t old_size = col.size();
159
6.17k
        col.resize(old_size + 1);
160
6.17k
        (reinterpret_cast<Data*>(col.get_data().data()) + old_size)->count =
161
6.17k
                AggregateFunctionCount::data(place).count;
162
6.17k
    }
163
164
11.3k
    MutableColumnPtr create_serialize_column() const override {
165
11.3k
        return ColumnFixedLengthObject::create(sizeof(Data));
166
11.3k
    }
167
168
11.4k
    DataTypePtr get_serialized_type() const override {
169
11.4k
        return std::make_shared<DataTypeFixedLengthObject>();
170
11.4k
    }
171
172
    void add_range_single_place(int64_t partition_start, int64_t partition_end, int64_t frame_start,
173
                                int64_t frame_end, AggregateDataPtr place, const IColumn** columns,
174
                                Arena& arena, UInt8* use_null_result,
175
264
                                UInt8* could_use_previous_result) const override {
176
264
        frame_start = std::max<int64_t>(frame_start, partition_start);
177
264
        frame_end = std::min<int64_t>(frame_end, partition_end);
178
264
        if (frame_start >= frame_end) {
179
19
            if (!*could_use_previous_result) {
180
2
                *use_null_result = true;
181
2
            }
182
245
        } else {
183
245
            AggregateFunctionCount::data(place).count += frame_end - frame_start;
184
245
            *use_null_result = false;
185
245
            *could_use_previous_result = true;
186
245
        }
187
264
    }
188
};
189
190
// Used for unary count(nullable_expr). SQL count(expr) counts non-NULL values.
191
class AggregateFunctionCountNotNullUnary final
192
        : public IAggregateFunctionDataHelper<AggregateFunctionCountData,
193
                                              AggregateFunctionCountNotNullUnary> {
194
public:
195
    AggregateFunctionCountNotNullUnary(const DataTypes& argument_types_)
196
13.9k
            : IAggregateFunctionDataHelper(argument_types_) {}
197
198
213
    String get_name() const override { return "count"; }
199
200
47.2k
    DataTypePtr get_return_type() const override { return std::make_shared<DataTypeInt64>(); }
201
202
0
    bool is_trivial() const override { return true; }
203
204
    void add(AggregateDataPtr __restrict place, const IColumn** columns, ssize_t row_num,
205
1.55M
             Arena&) const override {
206
1.55M
        data(place).count +=
207
1.55M
                !assert_cast<const ColumnNullable&, TypeCheckOnRelease::DISABLE>(*columns[0])
208
1.55M
                         .is_null_at(row_num);
209
1.55M
    }
210
211
    void add_batch_single_place(size_t batch_size, AggregateDataPtr place, const IColumn** columns,
212
14.5k
                                Arena&) const override {
213
14.5k
        const auto& nullable_column =
214
14.5k
                assert_cast<const ColumnNullable&, TypeCheckOnRelease::DISABLE>(*columns[0]);
215
14.5k
        const auto& null_map = nullable_column.get_null_map_data();
216
14.5k
        DCHECK_LE(batch_size, null_map.size());
217
14.5k
        if (!nullable_column.has_null(0, batch_size)) {
218
11.7k
            data(place).count += batch_size;
219
11.7k
            return;
220
11.7k
        }
221
2.72k
        data(place).count +=
222
2.72k
                simd::count_zero_num(reinterpret_cast<const int8_t*>(null_map.data()), batch_size);
223
2.72k
    }
224
225
253
    void reset(AggregateDataPtr place) const override { data(place).count = 0; }
226
227
    void merge(AggregateDataPtr __restrict place, ConstAggregateDataPtr rhs,
228
92.5k
               Arena&) const override {
229
92.5k
        data(place).count += data(rhs).count;
230
92.5k
    }
231
232
320
    void serialize(ConstAggregateDataPtr __restrict place, BufferWritable& buf) const override {
233
320
        buf.write_var_uint(data(place).count);
234
320
    }
235
236
    void deserialize(AggregateDataPtr __restrict place, BufferReadable& buf,
237
320
                     Arena&) const override {
238
320
        buf.read_var_uint(data(place).count);
239
320
    }
240
241
102k
    void insert_result_into(ConstAggregateDataPtr __restrict place, IColumn& to) const override {
242
102k
        if (is_column_nullable(to)) {
243
1
            auto& null_column = assert_cast<ColumnNullable&, TypeCheckOnRelease::DISABLE>(to);
244
1
            null_column.get_null_map_data().push_back(0);
245
1
            assert_cast<ColumnInt64&, TypeCheckOnRelease::DISABLE>(null_column.get_nested_column())
246
1
                    .get_data()
247
1
                    .push_back(data(place).count);
248
102k
        } else {
249
102k
            assert_cast<ColumnInt64&, TypeCheckOnRelease::DISABLE>(to).get_data().push_back(
250
102k
                    data(place).count);
251
102k
        }
252
102k
    }
253
254
13.8k
    void check_result_column_type(const IColumn& to) const override {
255
13.8k
        if (const auto* null_column = check_and_get_column<ColumnNullable>(to)) {
256
1
            IAggregateFunction::check_result_column_type(null_column->get_nested_column());
257
1
            return;
258
1
        }
259
13.8k
        IAggregateFunction::check_result_column_type(to);
260
13.8k
    }
261
262
    void serialize_to_column(const std::vector<AggregateDataPtr>& places, size_t offset,
263
4.06k
                             MutableColumnPtr& dst, const size_t num_rows) const override {
264
4.06k
        auto& col = assert_cast<ColumnFixedLengthObject&>(*dst);
265
18.4E
        DCHECK(col.item_size() == sizeof(Data))
266
18.4E
                << "size is not equal: " << col.item_size() << " " << sizeof(Data);
267
4.06k
        col.resize(num_rows);
268
4.06k
        auto* data = col.get_data().data();
269
246k
        for (size_t i = 0; i != num_rows; ++i) {
270
242k
            *reinterpret_cast<Data*>(&data[sizeof(Data) * i]) =
271
242k
                    *reinterpret_cast<Data*>(places[i] + offset);
272
242k
        }
273
4.06k
    }
274
275
    void streaming_agg_serialize_to_column(const IColumn** columns, MutableColumnPtr& dst,
276
70
                                           const size_t num_rows, Arena&) const override {
277
70
        auto& col = assert_cast<ColumnFixedLengthObject&>(*dst);
278
70
        DCHECK(col.item_size() == sizeof(Data))
279
0
                << "size is not equal: " << col.item_size() << " " << sizeof(Data);
280
70
        col.resize(num_rows);
281
70
        auto& data = col.get_data();
282
70
        const ColumnNullable& input_col = assert_cast<const ColumnNullable&>(*columns[0]);
283
211
        for (size_t i = 0; i < num_rows; i++) {
284
141
            auto& state = *reinterpret_cast<Data*>(&data[sizeof(Data) * i]);
285
141
            state.count = !input_col.is_null_at(i);
286
141
        }
287
70
    }
288
289
    void deserialize_and_merge_from_column_range(AggregateDataPtr __restrict place,
290
                                                 const IColumn& column, size_t begin, size_t end,
291
5.89k
                                                 Arena&) const override {
292
18.4E
        DCHECK(end <= column.size() && begin <= end)
293
18.4E
                << ", begin:" << begin << ", end:" << end << ", column.size():" << column.size();
294
5.89k
        auto& col = assert_cast<const ColumnFixedLengthObject&>(column);
295
5.89k
        auto* data = reinterpret_cast<const Data*>(col.get_data().data());
296
86.2k
        for (size_t i = begin; i <= end; ++i) {
297
80.3k
            doris::AggregateFunctionCountNotNullUnary::data(place).count += data[i].count;
298
80.3k
        }
299
5.89k
    }
300
301
    void deserialize_and_merge_vec(const AggregateDataPtr* places, size_t offset,
302
                                   AggregateDataPtr rhs, const IColumn* column, Arena& arena,
303
2.49k
                                   const size_t num_rows) const override {
304
2.49k
        const auto& col = assert_cast<const ColumnFixedLengthObject&>(*column);
305
2.49k
        const auto* data = col.get_data().data();
306
2.49k
        this->merge_vec(places, offset, AggregateDataPtr(data), arena, num_rows);
307
2.49k
    }
308
309
    void deserialize_and_merge_vec_selected(const AggregateDataPtr* places, size_t offset,
310
                                            AggregateDataPtr rhs, const IColumn* column,
311
2
                                            Arena& arena, const size_t num_rows) const override {
312
2
        const auto& col = assert_cast<const ColumnFixedLengthObject&>(*column);
313
2
        const auto* data = col.get_data().data();
314
2
        this->merge_vec_selected(places, offset, AggregateDataPtr(data), arena, num_rows);
315
2
    }
316
317
    void serialize_without_key_to_column(ConstAggregateDataPtr __restrict place,
318
4.93k
                                         IColumn& to) const override {
319
4.93k
        auto& col = assert_cast<ColumnFixedLengthObject&>(to);
320
18.4E
        DCHECK(col.item_size() == sizeof(Data))
321
18.4E
                << "size is not equal: " << col.item_size() << " " << sizeof(Data);
322
4.93k
        size_t old_size = col.size();
323
4.93k
        col.resize(old_size + 1);
324
4.93k
        (reinterpret_cast<Data*>(col.get_data().data()) + old_size)->count =
325
4.93k
                AggregateFunctionCountNotNullUnary::data(place).count;
326
4.93k
    }
327
328
9.34k
    MutableColumnPtr create_serialize_column() const override {
329
9.34k
        return ColumnFixedLengthObject::create(sizeof(Data));
330
9.34k
    }
331
332
9.58k
    DataTypePtr get_serialized_type() const override {
333
9.58k
        return std::make_shared<DataTypeFixedLengthObject>();
334
9.58k
    }
335
336
    void add_range_single_place(int64_t partition_start, int64_t partition_end, int64_t frame_start,
337
                                int64_t frame_end, AggregateDataPtr place, const IColumn** columns,
338
                                Arena& arena, UInt8* use_null_result,
339
337
                                UInt8* could_use_previous_result) const override {
340
337
        frame_start = std::max<int64_t>(frame_start, partition_start);
341
337
        frame_end = std::min<int64_t>(frame_end, partition_end);
342
337
        if (frame_start >= frame_end) {
343
12
            if (!*could_use_previous_result) {
344
0
                *use_null_result = true;
345
0
            }
346
325
        } else {
347
325
            const auto& nullable_column =
348
325
                    assert_cast<const ColumnNullable&, TypeCheckOnRelease::DISABLE>(*columns[0]);
349
325
            size_t count = 0;
350
325
            if (nullable_column.has_null()) {
351
358
                for (int64_t i = frame_start; i < frame_end; ++i) {
352
233
                    if (!nullable_column.is_null_at(i)) {
353
166
                        ++count;
354
166
                    }
355
233
                }
356
200
            } else {
357
200
                count = frame_end - frame_start;
358
200
            }
359
325
            *use_null_result = false;
360
325
            *could_use_previous_result = true;
361
325
            AggregateFunctionCountNotNullUnary::data(place).count += count;
362
325
        }
363
337
    }
364
};
365
366
} // namespace doris