Coverage Report

Created: 2026-03-12 14:13

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
be/src/exprs/function/function_convert_tz.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 <cctz/time_zone.h>
19
20
#include <cstddef>
21
#include <cstdint>
22
#include <memory>
23
#include <string>
24
#include <utility>
25
26
#include "common/status.h"
27
#include "core/assert_cast.h"
28
#include "core/binary_cast.hpp"
29
#include "core/block/block.h"
30
#include "core/block/column_numbers.h"
31
#include "core/block/column_with_type_and_name.h"
32
#include "core/column/column.h"
33
#include "core/column/column_const.h"
34
#include "core/column/column_nullable.h"
35
#include "core/column/column_string.h"
36
#include "core/column/column_vector.h"
37
#include "core/data_type/data_type.h"
38
#include "core/data_type/data_type_date.h"
39
#include "core/data_type/data_type_date_or_datetime_v2.h"
40
#include "core/data_type/data_type_date_time.h"
41
#include "core/data_type/data_type_nullable.h"
42
#include "core/data_type/data_type_string.h"
43
#include "core/data_type/define_primitive_type.h"
44
#include "core/data_type/primitive_type.h"
45
#include "core/string_ref.h"
46
#include "core/types.h"
47
#include "core/value/vdatetime_value.h"
48
#include "exec/common/util.hpp"
49
#include "exprs/aggregate/aggregate_function.h"
50
#include "exprs/function/datetime_errors.h"
51
#include "exprs/function/function.h"
52
#include "exprs/function/function_helpers.h"
53
#include "exprs/function/simple_function_factory.h"
54
#include "exprs/function_context.h"
55
#include "util/timezone_utils.h"
56
57
namespace doris {
58
#include "common/compile_check_begin.h"
59
60
struct ConvertTzState {
61
    bool use_state = false;
62
    bool is_valid = false;
63
    cctz::time_zone from_tz;
64
    cctz::time_zone to_tz;
65
};
66
67
class FunctionConvertTZ : public IFunction {
68
    constexpr static PrimitiveType PType = PrimitiveType::TYPE_DATETIMEV2;
69
    using DateValueType = PrimitiveTypeTraits<PType>::CppType;
70
    using ColumnType = PrimitiveTypeTraits<PType>::ColumnType;
71
72
public:
73
    static constexpr auto name = "convert_tz";
74
75
87
    static FunctionPtr create() { return std::make_shared<FunctionConvertTZ>(); }
76
77
1
    String get_name() const override { return name; }
78
79
78
    size_t get_number_of_arguments() const override { return 3; }
80
81
78
    DataTypePtr get_return_type_impl(const DataTypes& arguments) const override {
82
78
        return have_nullable(arguments) ? make_nullable(std::make_shared<DataTypeDateTimeV2>())
83
78
                                        : std::make_shared<DataTypeDateTimeV2>();
84
78
    }
85
86
    // default value of timezone is invalid, should skip to avoid wrong exception
87
173
    bool use_default_implementation_for_nulls() const override { return false; }
88
89
363
    Status open(FunctionContext* context, FunctionContext::FunctionStateScope scope) override {
90
363
        if (scope == FunctionContext::THREAD_LOCAL) {
91
285
            return Status::OK();
92
285
        }
93
78
        std::shared_ptr<ConvertTzState> state = std::make_shared<ConvertTzState>();
94
95
78
        context->set_function_state(scope, state);
96
78
        DCHECK_EQ(context->get_num_args(), 3);
97
78
        const auto* const_from_tz = context->get_constant_col(1);
98
78
        const auto* const_to_tz = context->get_constant_col(2);
99
100
        // ConvertTzState is used only when both the second and third parameters are constants
101
78
        if (const_from_tz != nullptr && const_to_tz != nullptr) {
102
44
            state->use_state = true;
103
44
            init_convert_tz_state(state, const_from_tz, const_to_tz);
104
44
        } else {
105
34
            state->use_state = false;
106
34
        }
107
108
78
        return IFunction::open(context, scope);
109
363
    }
110
111
    void init_convert_tz_state(std::shared_ptr<ConvertTzState> state,
112
                               const ColumnPtrWrapper* const_from_tz,
113
44
                               const ColumnPtrWrapper* const_to_tz) {
114
44
        auto const_data_from_tz = const_from_tz->column_ptr->get_data_at(0);
115
44
        auto const_data_to_tz = const_to_tz->column_ptr->get_data_at(0);
116
117
        // from_tz and to_tz must both be non-null.
118
44
        if (const_data_from_tz.data == nullptr || const_data_to_tz.data == nullptr) {
119
0
            state->is_valid = false;
120
0
            return;
121
0
        }
122
123
44
        auto from_tz_name = const_data_from_tz.to_string();
124
44
        auto to_tz_name = const_data_to_tz.to_string();
125
126
44
        if (!TimezoneUtils::find_cctz_time_zone(from_tz_name, state->from_tz)) [[unlikely]] {
127
2
            throw Exception(ErrorCode::INVALID_ARGUMENT, "Operation {} invalid timezone: {}", name,
128
2
                            from_tz_name);
129
2
        }
130
42
        if (!TimezoneUtils::find_cctz_time_zone(to_tz_name, state->to_tz)) [[unlikely]] {
131
2
            throw Exception(ErrorCode::INVALID_ARGUMENT, "Operation {} invalid timezone: {}", name,
132
2
                            to_tz_name);
133
2
        }
134
40
        state->is_valid = true;
135
40
    }
136
137
    Status execute_impl(FunctionContext* context, Block& block, const ColumnNumbers& arguments,
138
95
                        uint32_t result, size_t input_rows_count) const override {
139
95
        auto* convert_tz_state = reinterpret_cast<ConvertTzState*>(
140
95
                context->get_function_state(FunctionContext::FRAGMENT_LOCAL));
141
95
        if (!convert_tz_state) {
142
0
            return Status::RuntimeError(
143
0
                    "funciton context for function '{}' must have ConvertTzState;", get_name());
144
0
        }
145
146
95
        auto result_null_map_column = ColumnUInt8::create(input_rows_count, 0);
147
95
        NullMap& result_null_map = assert_cast<ColumnUInt8&>(*result_null_map_column).get_data();
148
149
95
        ColumnPtr argument_columns[3];
150
95
        bool col_const[3];
151
152
        // calculate result null map and col_const
153
380
        for (int i = 0; i < 3; ++i) {
154
285
            ColumnPtr& col = block.get_by_position(arguments[i]).column;
155
285
            col_const[i] = is_column_const(*col);
156
285
            const NullMap* null_map = VectorizedUtils::get_null_map(col);
157
285
            if (null_map) {
158
141
                VectorizedUtils::update_null_map(result_null_map, *null_map, col_const[i]);
159
141
            }
160
285
        }
161
162
        // Extract nested columns from const(nullable) wrappers
163
95
        argument_columns[0] = col_const[0] ? static_cast<const ColumnConst&>(
164
1
                                                     *block.get_by_position(arguments[0]).column)
165
1
                                                     .convert_to_full_column()
166
95
                                           : block.get_by_position(arguments[0]).column;
167
95
        argument_columns[0] = remove_nullable(argument_columns[0]);
168
95
        default_preprocess_parameter_columns(argument_columns, col_const, {1, 2}, block, arguments);
169
95
        argument_columns[1] = remove_nullable(argument_columns[1]);
170
95
        argument_columns[2] = remove_nullable(argument_columns[2]);
171
172
95
        auto result_column = ColumnType::create();
173
95
        if (convert_tz_state->use_state) {
174
            // ignore argument columns, use cached timezone input in state
175
52
            execute_tz_const_with_state(convert_tz_state,
176
52
                                        assert_cast<const ColumnType*>(argument_columns[0].get()),
177
52
                                        assert_cast<ColumnType*>(result_column.get()),
178
52
                                        result_null_map, input_rows_count);
179
52
        } else if (col_const[1] && col_const[2]) {
180
            // arguments are const
181
0
            execute_tz_const(context, assert_cast<const ColumnType*>(argument_columns[0].get()),
182
0
                             assert_cast<const ColumnString*>(argument_columns[1].get()),
183
0
                             assert_cast<const ColumnString*>(argument_columns[2].get()),
184
0
                             assert_cast<ColumnType*>(result_column.get()), result_null_map,
185
0
                             input_rows_count);
186
43
        } else {
187
43
            _execute(context, assert_cast<const ColumnType*>(argument_columns[0].get()),
188
43
                     assert_cast<const ColumnString*>(argument_columns[1].get()),
189
43
                     assert_cast<const ColumnString*>(argument_columns[2].get()),
190
43
                     assert_cast<ColumnType*>(result_column.get()), result_null_map,
191
43
                     input_rows_count);
192
43
        } //if const
193
194
95
        if (block.get_data_type(result)->is_nullable()) {
195
84
            block.get_by_position(result).column = ColumnNullable::create(
196
84
                    std::move(result_column), std::move(result_null_map_column));
197
84
        } else {
198
11
            block.get_by_position(result).column = std::move(result_column);
199
11
        }
200
95
        return Status::OK();
201
95
    }
202
203
private:
204
    static void _execute(FunctionContext* context, const ColumnType* date_column,
205
                         const ColumnString* from_tz_column, const ColumnString* to_tz_column,
206
                         ColumnType* result_column, NullMap& result_null_map,
207
43
                         size_t input_rows_count) {
208
169
        for (size_t i = 0; i < input_rows_count; i++) {
209
126
            if (result_null_map[i]) {
210
0
                result_column->insert_default();
211
0
                continue;
212
0
            }
213
126
            auto from_tz = from_tz_column->get_data_at(i).to_string();
214
126
            auto to_tz = to_tz_column->get_data_at(i).to_string();
215
126
            execute_inner_loop(date_column, from_tz, to_tz, result_column, result_null_map, i);
216
126
        }
217
43
    }
218
219
    static void execute_tz_const_with_state(ConvertTzState* convert_tz_state,
220
                                            const ColumnType* date_column,
221
                                            ColumnType* result_column, NullMap& result_null_map,
222
52
                                            size_t input_rows_count) {
223
52
        cctz::time_zone& from_tz = convert_tz_state->from_tz;
224
52
        cctz::time_zone& to_tz = convert_tz_state->to_tz;
225
52
        auto push_null = [&](size_t row) {
226
0
            result_null_map[row] = true;
227
0
            result_column->insert_default();
228
0
        };
229
        // state isn't valid means there's NULL in timezone input. so return null rather than exception
230
52
        if (!convert_tz_state->is_valid) [[unlikely]] {
231
            // If an invalid timezone is present, return null
232
0
            for (size_t i = 0; i < input_rows_count; i++) {
233
0
                push_null(i);
234
0
            }
235
0
            return;
236
0
        }
237
157
        for (size_t i = 0; i < input_rows_count; i++) {
238
105
            if (result_null_map[i]) {
239
2
                result_column->insert_default();
240
2
                continue;
241
2
            }
242
243
103
            DateValueType ts_value = date_column->get_element(i);
244
103
            DateValueType ts_value2;
245
246
103
            std::pair<int64_t, int64_t> timestamp;
247
103
            ts_value.unix_timestamp(&timestamp, from_tz);
248
103
            ts_value2.from_unixtime(timestamp, to_tz);
249
250
103
            if (!ts_value2.is_valid_date()) [[unlikely]] {
251
1
                throw_out_of_bound_convert_tz<DateValueType>(date_column->get_element(i),
252
1
                                                             from_tz.name(), to_tz.name());
253
1
            }
254
255
103
            result_column->insert(Field::create_field<TYPE_DATETIMEV2>(ts_value2));
256
103
        }
257
52
    }
258
259
    static void execute_tz_const(FunctionContext* context, const ColumnType* date_column,
260
                                 const ColumnString* from_tz_column,
261
                                 const ColumnString* to_tz_column, ColumnType* result_column,
262
0
                                 NullMap& result_null_map, size_t input_rows_count) {
263
0
        auto from_tz = from_tz_column->get_data_at(0).to_string();
264
0
        auto to_tz = to_tz_column->get_data_at(0).to_string();
265
0
        cctz::time_zone from_zone, to_zone;
266
0
        if (!TimezoneUtils::find_cctz_time_zone(from_tz, from_zone)) [[unlikely]] {
267
0
            throw Exception(ErrorCode::INVALID_ARGUMENT, "Operation {} invalid timezone: {}", name,
268
0
                            from_tz);
269
0
        }
270
0
        if (!TimezoneUtils::find_cctz_time_zone(to_tz, to_zone)) [[unlikely]] {
271
0
            throw Exception(ErrorCode::INVALID_ARGUMENT, "Operation {} invalid timezone: {}", name,
272
0
                            to_tz);
273
0
        }
274
0
        for (size_t i = 0; i < input_rows_count; i++) {
275
0
            if (result_null_map[i]) {
276
0
                result_column->insert_default();
277
0
                continue;
278
0
            }
279
0
            execute_inner_loop(date_column, from_tz, to_tz, result_column, result_null_map, i);
280
0
        }
281
0
    }
282
283
    static void execute_inner_loop(const ColumnType* date_column, const std::string& from_tz_name,
284
                                   const std::string& to_tz_name, ColumnType* result_column,
285
126
                                   NullMap& result_null_map, const size_t index_now) {
286
126
        DateValueType ts_value = date_column->get_element(index_now);
287
126
        cctz::time_zone from_tz {}, to_tz {};
288
126
        DateValueType ts_value2;
289
290
126
        if (!TimezoneUtils::find_cctz_time_zone(from_tz_name, from_tz)) [[unlikely]] {
291
0
            throw Exception(ErrorCode::INVALID_ARGUMENT, "Operation {} invalid timezone: {}", name,
292
0
                            from_tz_name);
293
0
        }
294
126
        if (!TimezoneUtils::find_cctz_time_zone(to_tz_name, to_tz)) [[unlikely]] {
295
0
            throw Exception(ErrorCode::INVALID_ARGUMENT, "Operation {} invalid timezone: {}", name,
296
0
                            to_tz_name);
297
0
        }
298
299
126
        std::pair<int64_t, int64_t> timestamp;
300
126
        ts_value.unix_timestamp(&timestamp, from_tz);
301
126
        ts_value2.from_unixtime(timestamp, to_tz);
302
303
126
        if (!ts_value2.is_valid_date()) [[unlikely]] {
304
0
            throw_out_of_bound_convert_tz<DateValueType>(date_column->get_element(index_now),
305
0
                                                         from_tz.name(), to_tz.name());
306
0
        }
307
308
126
        result_column->insert(Field::create_field<TYPE_DATETIMEV2>(ts_value2));
309
126
    }
310
};
311
312
8
void register_function_convert_tz(SimpleFunctionFactory& factory) {
313
8
    factory.register_function<FunctionConvertTZ>();
314
8
}
315
316
} // namespace doris
317
318
#include "common/compile_check_end.h"