be/src/exprs/function/function_hex.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 <stddef.h> |
19 | | #include <stdint.h> |
20 | | |
21 | | #include <algorithm> |
22 | | #include <boost/iterator/iterator_facade.hpp> |
23 | | #include <memory> |
24 | | #include <string> |
25 | | #include <string_view> |
26 | | #include <utility> |
27 | | |
28 | | #include "common/cast_set.h" |
29 | | #include "common/status.h" |
30 | | #include "core/block/block.h" |
31 | | #include "core/block/column_numbers.h" |
32 | | #include "core/block/column_with_type_and_name.h" |
33 | | #include "core/column/column.h" |
34 | | #include "core/column/column_complex.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_hll.h" |
39 | | #include "core/data_type/data_type_number.h" |
40 | | #include "core/data_type/data_type_string.h" |
41 | | #include "core/types.h" |
42 | | #include "exec/common/stringop_substring.h" |
43 | | #include "exprs/aggregate/aggregate_function.h" |
44 | | #include "exprs/function/function.h" |
45 | | #include "exprs/function/simple_function_factory.h" |
46 | | #include "exprs/function/string_hex_util.h" |
47 | | |
48 | | namespace doris { |
49 | | #include "common/compile_check_begin.h" |
50 | | class FunctionContext; |
51 | | } // namespace doris |
52 | | |
53 | | namespace doris { |
54 | | template <typename Impl> |
55 | | class FunctionHexVariadic : public IFunction { |
56 | | public: |
57 | | static constexpr auto name = "hex"; |
58 | | |
59 | 271 | static FunctionPtr create() { return std::make_shared<FunctionHexVariadic>(); }_ZN5doris19FunctionHexVariadicINS_13HexStringImplEE6createEv Line | Count | Source | 59 | 201 | static FunctionPtr create() { return std::make_shared<FunctionHexVariadic>(); } |
_ZN5doris19FunctionHexVariadicINS_10HexIntImplEE6createEv Line | Count | Source | 59 | 61 | static FunctionPtr create() { return std::make_shared<FunctionHexVariadic>(); } |
_ZN5doris19FunctionHexVariadicINS_10HexHLLImplEE6createEv Line | Count | Source | 59 | 9 | static FunctionPtr create() { return std::make_shared<FunctionHexVariadic>(); } |
|
60 | | |
61 | 3 | String get_name() const override { return name; }_ZNK5doris19FunctionHexVariadicINS_13HexStringImplEE8get_nameB5cxx11Ev Line | Count | Source | 61 | 1 | String get_name() const override { return name; } |
_ZNK5doris19FunctionHexVariadicINS_10HexIntImplEE8get_nameB5cxx11Ev Line | Count | Source | 61 | 1 | String get_name() const override { return name; } |
_ZNK5doris19FunctionHexVariadicINS_10HexHLLImplEE8get_nameB5cxx11Ev Line | Count | Source | 61 | 1 | String get_name() const override { return name; } |
|
62 | | |
63 | 244 | size_t get_number_of_arguments() const override { return 1; }_ZNK5doris19FunctionHexVariadicINS_13HexStringImplEE23get_number_of_argumentsEv Line | Count | Source | 63 | 192 | size_t get_number_of_arguments() const override { return 1; } |
_ZNK5doris19FunctionHexVariadicINS_10HexIntImplEE23get_number_of_argumentsEv Line | Count | Source | 63 | 52 | size_t get_number_of_arguments() const override { return 1; } |
Unexecuted instantiation: _ZNK5doris19FunctionHexVariadicINS_10HexHLLImplEE23get_number_of_argumentsEv |
64 | | |
65 | 244 | DataTypePtr get_return_type_impl(const DataTypes& arguments) const override { |
66 | 244 | return std::make_shared<DataTypeString>(); |
67 | 244 | } _ZNK5doris19FunctionHexVariadicINS_13HexStringImplEE20get_return_type_implERKSt6vectorISt10shared_ptrIKNS_9IDataTypeEESaIS7_EE Line | Count | Source | 65 | 192 | DataTypePtr get_return_type_impl(const DataTypes& arguments) const override { | 66 | 192 | return std::make_shared<DataTypeString>(); | 67 | 192 | } |
_ZNK5doris19FunctionHexVariadicINS_10HexIntImplEE20get_return_type_implERKSt6vectorISt10shared_ptrIKNS_9IDataTypeEESaIS7_EE Line | Count | Source | 65 | 52 | DataTypePtr get_return_type_impl(const DataTypes& arguments) const override { | 66 | 52 | return std::make_shared<DataTypeString>(); | 67 | 52 | } |
Unexecuted instantiation: _ZNK5doris19FunctionHexVariadicINS_10HexHLLImplEE20get_return_type_implERKSt6vectorISt10shared_ptrIKNS_9IDataTypeEESaIS7_EE |
68 | | |
69 | 24 | DataTypes get_variadic_argument_types_impl() const override { |
70 | 24 | return Impl::get_variadic_argument_types(); |
71 | 24 | } _ZNK5doris19FunctionHexVariadicINS_13HexStringImplEE32get_variadic_argument_types_implEv Line | Count | Source | 69 | 8 | DataTypes get_variadic_argument_types_impl() const override { | 70 | 8 | return Impl::get_variadic_argument_types(); | 71 | 8 | } |
_ZNK5doris19FunctionHexVariadicINS_10HexIntImplEE32get_variadic_argument_types_implEv Line | Count | Source | 69 | 8 | DataTypes get_variadic_argument_types_impl() const override { | 70 | 8 | return Impl::get_variadic_argument_types(); | 71 | 8 | } |
_ZNK5doris19FunctionHexVariadicINS_10HexHLLImplEE32get_variadic_argument_types_implEv Line | Count | Source | 69 | 8 | DataTypes get_variadic_argument_types_impl() const override { | 70 | 8 | return Impl::get_variadic_argument_types(); | 71 | 8 | } |
|
72 | | |
73 | | Status execute_impl(FunctionContext* context, Block& block, const ColumnNumbers& arguments, |
74 | 293 | uint32_t result, size_t input_rows_count) const override { |
75 | 293 | ColumnPtr& argument_column = block.get_by_position(arguments[0]).column; |
76 | | |
77 | 293 | auto result_data_column = ColumnString::create(); |
78 | 293 | auto& result_data = result_data_column->get_chars(); |
79 | 293 | auto& result_offset = result_data_column->get_offsets(); |
80 | | |
81 | 293 | RETURN_IF_ERROR( |
82 | 293 | Impl::vector(argument_column, input_rows_count, result_data, result_offset)); |
83 | 293 | block.replace_by_position(result, std::move(result_data_column)); |
84 | 293 | return Status::OK(); |
85 | 293 | } _ZNK5doris19FunctionHexVariadicINS_13HexStringImplEE12execute_implEPNS_15FunctionContextERNS_5BlockERKSt6vectorIjSaIjEEjm Line | Count | Source | 74 | 217 | uint32_t result, size_t input_rows_count) const override { | 75 | 217 | ColumnPtr& argument_column = block.get_by_position(arguments[0]).column; | 76 | | | 77 | 217 | auto result_data_column = ColumnString::create(); | 78 | 217 | auto& result_data = result_data_column->get_chars(); | 79 | 217 | auto& result_offset = result_data_column->get_offsets(); | 80 | | | 81 | 217 | RETURN_IF_ERROR( | 82 | 217 | Impl::vector(argument_column, input_rows_count, result_data, result_offset)); | 83 | 217 | block.replace_by_position(result, std::move(result_data_column)); | 84 | 217 | return Status::OK(); | 85 | 217 | } |
_ZNK5doris19FunctionHexVariadicINS_10HexIntImplEE12execute_implEPNS_15FunctionContextERNS_5BlockERKSt6vectorIjSaIjEEjm Line | Count | Source | 74 | 76 | uint32_t result, size_t input_rows_count) const override { | 75 | 76 | ColumnPtr& argument_column = block.get_by_position(arguments[0]).column; | 76 | | | 77 | 76 | auto result_data_column = ColumnString::create(); | 78 | 76 | auto& result_data = result_data_column->get_chars(); | 79 | 76 | auto& result_offset = result_data_column->get_offsets(); | 80 | | | 81 | 76 | RETURN_IF_ERROR( | 82 | 76 | Impl::vector(argument_column, input_rows_count, result_data, result_offset)); | 83 | 76 | block.replace_by_position(result, std::move(result_data_column)); | 84 | 76 | return Status::OK(); | 85 | 76 | } |
Unexecuted instantiation: _ZNK5doris19FunctionHexVariadicINS_10HexHLLImplEE12execute_implEPNS_15FunctionContextERNS_5BlockERKSt6vectorIjSaIjEEjm |
86 | | }; |
87 | | |
88 | | struct HexStringImpl { |
89 | 8 | static DataTypes get_variadic_argument_types() { return {std::make_shared<DataTypeString>()}; } |
90 | | |
91 | | static Status vector(ColumnPtr argument_column, size_t input_rows_count, |
92 | 217 | ColumnString::Chars& dst_data, ColumnString::Offsets& dst_offsets) { |
93 | 217 | const auto* str_col = check_and_get_column<ColumnString>(argument_column.get()); |
94 | 217 | const auto& data = str_col->get_chars(); |
95 | 217 | const auto& offsets = str_col->get_offsets(); |
96 | 217 | dst_offsets.resize(input_rows_count); |
97 | 217 | dst_data.resize(data.size() * 2); |
98 | | |
99 | 217 | size_t offset = 0; |
100 | 217 | auto* dst_data_ptr = dst_data.data(); |
101 | 1.26k | for (int i = 0; i < input_rows_count; ++i) { |
102 | 1.04k | const auto* source = reinterpret_cast<const unsigned char*>(&data[offsets[i - 1]]); |
103 | 1.04k | size_t srclen = offsets[i] - offsets[i - 1]; |
104 | 1.04k | string_hex::hex_encode(source, srclen, dst_data_ptr, offset); |
105 | 1.04k | dst_offsets[i] = cast_set<uint32_t>(offset); |
106 | 1.04k | } |
107 | 217 | return Status::OK(); |
108 | 217 | } |
109 | | }; |
110 | | |
111 | | struct HexIntImpl { |
112 | 8 | static DataTypes get_variadic_argument_types() { return {std::make_shared<DataTypeInt64>()}; } |
113 | | |
114 | 281 | static std::string_view hex(uint64_t num, char* ans) { |
115 | 281 | static constexpr auto hex_table = "0123456789ABCDEF"; |
116 | | // uint64_t max value 0xFFFFFFFFFFFFFFFF , 16 'F' |
117 | 281 | if (num == 0) { |
118 | 114 | return {hex_table, 1}; |
119 | 114 | } |
120 | | |
121 | 167 | int i = 0; |
122 | 1.37k | while (num) { |
123 | 1.20k | ans[i++] = hex_table[num & 15]; |
124 | 1.20k | num = num >> 4; |
125 | 1.20k | } |
126 | 167 | ans[i] = '\0'; |
127 | | |
128 | | // reverse |
129 | 790 | for (int k = 0, j = i - 1; k <= j && k <= 16; k++, j--) { |
130 | 623 | char tmp = ans[j]; |
131 | 623 | ans[j] = ans[k]; |
132 | 623 | ans[k] = tmp; |
133 | 623 | } |
134 | | |
135 | 167 | return {ans, static_cast<size_t>(i)}; |
136 | 281 | } |
137 | | |
138 | | static Status vector(ColumnPtr argument_column, size_t input_rows_count, |
139 | 76 | ColumnString::Chars& res_data, ColumnString::Offsets& res_offsets) { |
140 | 76 | const auto* str_col = check_and_get_column<ColumnInt64>(argument_column.get()); |
141 | 76 | const auto& data = str_col->get_data(); |
142 | | |
143 | 76 | res_offsets.resize(input_rows_count); |
144 | 76 | char ans[17]; |
145 | 357 | for (size_t i = 0; i < input_rows_count; ++i) { |
146 | 281 | StringOP::push_value_string(hex(data[i], ans), i, res_data, res_offsets); |
147 | 281 | } |
148 | 76 | return Status::OK(); |
149 | 76 | } |
150 | | }; |
151 | | |
152 | | struct HexHLLImpl { |
153 | 8 | static DataTypes get_variadic_argument_types() { return {std::make_shared<DataTypeHLL>()}; } |
154 | | |
155 | | static Status vector(ColumnPtr argument_column, size_t input_rows_count, |
156 | 0 | ColumnString::Chars& res_data, ColumnString::Offsets& res_offsets) { |
157 | 0 | const auto* str_col = check_and_get_column<ColumnHLL>(argument_column.get()); |
158 | 0 | const auto& hll_data = str_col->get_data(); |
159 | 0 | res_offsets.resize(input_rows_count); |
160 | 0 | size_t total_length = 0, offset = 0; |
161 | 0 | std::string hll_str; |
162 | 0 | unsigned char* dst_data_ptr = nullptr; |
163 | |
|
164 | 0 | for (size_t i = 0; i < input_rows_count; ++i) { |
165 | 0 | hll_str.resize(hll_data[i].max_serialized_size(), '0'); |
166 | 0 | size_t actual_size = hll_data[i].serialize((uint8_t*)hll_str.data()); |
167 | 0 | hll_str.resize(actual_size); |
168 | 0 | total_length += actual_size; |
169 | |
|
170 | 0 | res_data.resize(total_length * 2 + (i + 1)); |
171 | 0 | dst_data_ptr = res_data.data() + offset; |
172 | 0 | string_hex::hex_encode(reinterpret_cast<const unsigned char*>(hll_str.data()), |
173 | 0 | hll_str.length(), dst_data_ptr, offset); |
174 | 0 | res_offsets[i] = cast_set<uint32_t>(offset); |
175 | 0 | hll_str.clear(); |
176 | 0 | } |
177 | 0 | return Status::OK(); |
178 | 0 | } |
179 | | }; |
180 | | |
181 | 8 | void register_function_hex_variadic(SimpleFunctionFactory& factory) { |
182 | 8 | factory.register_function<FunctionHexVariadic<HexStringImpl>>(); |
183 | 8 | factory.register_function<FunctionHexVariadic<HexIntImpl>>(); |
184 | 8 | factory.register_function<FunctionHexVariadic<HexHLLImpl>>(); |
185 | 8 | } |
186 | | } // namespace doris |