Coverage Report

Created: 2025-06-11 12:48

/root/doris/be/src/vec/functions/like.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
18
#pragma once
19
20
#include <hs/hs_common.h>
21
#include <hs/hs_runtime.h>
22
#include <re2/re2.h>
23
#include <stddef.h>
24
#include <stdint.h>
25
26
#include <algorithm>
27
#include <boost/iterator/iterator_facade.hpp>
28
#include <functional>
29
#include <memory>
30
#include <string>
31
32
#include "common/status.h"
33
#include "runtime/define_primitive_type.h"
34
#include "runtime/string_search.hpp"
35
#include "udf/udf.h"
36
#include "vec/aggregate_functions/aggregate_function.h"
37
#include "vec/columns/column_string.h"
38
#include "vec/columns/predicate_column.h"
39
#include "vec/common/string_ref.h"
40
#include "vec/core/column_numbers.h"
41
#include "vec/core/types.h"
42
#include "vec/data_types/data_type_number.h"
43
#include "vec/functions/function.h"
44
45
namespace doris {
46
namespace vectorized {
47
class Block;
48
} // namespace vectorized
49
} // namespace doris
50
51
namespace doris::vectorized {
52
53
// TODO: replace with std::string_view when `LikeSearchState.substring_pattern` can
54
// construct from std::string_view.
55
struct LikeSearchState {
56
    char escape_char;
57
58
    /// Holds the string the StringRef points to and is set any time StringRef is
59
    /// used.
60
    std::string search_string;
61
62
    std::string pattern_str;
63
64
    /// Used for LIKE predicates if the pattern is a constant argument, and is either a
65
    /// constant string or has a constant string at the beginning or end of the pattern.
66
    /// This will be set in order to check for that pattern in the corresponding part of
67
    /// the string.
68
    StringRef search_string_sv;
69
70
    /// Used for LIKE predicates if the pattern is a constant argument and has a constant
71
    /// string in the middle of it. This will be use in order to check for the substring
72
    /// in the value.
73
    doris::StringSearch substring_pattern;
74
75
    /// Used for RLIKE and REGEXP predicates if the pattern is a constant argument.
76
    std::unique_ptr<re2::RE2> regex;
77
78
    template <typename Deleter, Deleter deleter>
79
    struct HyperscanDeleter {
80
        template <typename T>
81
10
        void operator()(T* ptr) const {
82
10
            deleter(ptr);
83
10
        }
_ZNK5doris10vectorized15LikeSearchState16HyperscanDeleterIPFiP11hs_databaseEXadL_Z16hs_free_databaseEEEclIS3_EEvPT_
Line
Count
Source
81
5
        void operator()(T* ptr) const {
82
5
            deleter(ptr);
83
5
        }
_ZNK5doris10vectorized15LikeSearchState16HyperscanDeleterIPFiP10hs_scratchEXadL_Z15hs_free_scratchEEEclIS3_EEvPT_
Line
Count
Source
81
5
        void operator()(T* ptr) const {
82
5
            deleter(ptr);
83
5
        }
84
    };
85
86
    // hyperscan compiled pattern database and scratch space, reused for performance
87
    std::unique_ptr<hs_database_t, HyperscanDeleter<decltype(&hs_free_database), &hs_free_database>>
88
            hs_database;
89
    std::unique_ptr<hs_scratch_t, HyperscanDeleter<decltype(&hs_free_scratch), &hs_free_scratch>>
90
            hs_scratch;
91
92
    // hyperscan match callback
93
    static int hs_match_handler(unsigned int /* from */,       // NOLINT
94
                                unsigned long long /* from */, // NOLINT
95
                                unsigned long long /* to */,   // NOLINT
96
3
                                unsigned int /* flags */, void* ctx) {
97
        // set result to 1 for matched row
98
3
        *((unsigned char*)ctx) = 1;
99
        /// return non-zero to indicate hyperscan stop after first matched
100
3
        return 1;
101
3
    }
102
103
18
    LikeSearchState() : escape_char('\\') {}
104
105
    Status clone(LikeSearchState& cloned);
106
107
12
    void set_search_string(const std::string& search_string_arg) {
108
12
        search_string = search_string_arg;
109
12
        search_string_sv = StringRef(search_string);
110
12
        substring_pattern.set_pattern(&search_string_sv);
111
12
    }
112
};
113
114
using LikeFn = std::function<doris::Status(LikeSearchState*, const ColumnString&, const StringRef&,
115
                                           ColumnUInt8::Container&)>;
116
117
using ScalarLikeFn = std::function<doris::Status(LikeSearchState*, const StringRef&,
118
                                                 const StringRef&, unsigned char*)>;
119
120
using VectorLikeFn = std::function<doris::Status(const ColumnString&, const ColumnString&,
121
                                                 ColumnUInt8::Container&)>;
122
123
struct LikeState {
124
    bool is_like_pattern;
125
    LikeSearchState search_state;
126
    LikeFn function;
127
    ScalarLikeFn scalar_function;
128
};
129
130
struct VectorPatternSearchState {
131
    MutableColumnPtr _search_strings;
132
    std::string _search_string;
133
    VectorLikeFn _vector_function;
134
    bool _pattern_matched;
135
136
    VectorPatternSearchState(VectorLikeFn vector_function)
137
            : _search_strings(ColumnString::create()),
138
              _vector_function(vector_function),
139
60
              _pattern_matched(true) {}
140
141
60
    virtual ~VectorPatternSearchState() = default;
142
143
    virtual void like_pattern_match(const std::string& pattern_str) = 0;
144
145
    virtual void regexp_pattern_match(const std::string& pattern_str) = 0;
146
};
147
148
using VPatternSearchStateSPtr = std::shared_ptr<VectorPatternSearchState>;
149
150
class FunctionLikeBase : public IFunction {
151
public:
152
20
    size_t get_number_of_arguments() const override { return 2; }
153
154
20
    DataTypePtr get_return_type_impl(const DataTypes& /*arguments*/) const override {
155
20
        return std::make_shared<DataTypeUInt8>();
156
20
    }
157
158
    Status execute_impl(FunctionContext* context, Block& block, const ColumnNumbers& arguments,
159
                        uint32_t result, size_t /*input_rows_count*/) const override;
160
161
    friend struct VectorAllpassSearchState;
162
    friend struct VectorEqualSearchState;
163
    friend struct VectorSubStringSearchState;
164
    friend struct VectorStartsWithSearchState;
165
    friend struct VectorEndsWithSearchState;
166
167
protected:
168
    Status vector_const(const ColumnString& values, const StringRef* pattern_val,
169
                        ColumnUInt8::Container& result, const LikeFn& function,
170
                        LikeSearchState* search_state) const;
171
172
    Status vector_non_const(const ColumnString& values, const ColumnString& patterns,
173
                            ColumnUInt8::Container& result, LikeState* state,
174
                            size_t input_rows_count) const;
175
176
    Status execute_substring(const ColumnString::Chars& values,
177
                             const ColumnString::Offsets& value_offsets,
178
                             ColumnUInt8::Container& result, LikeSearchState* search_state) const;
179
180
    template <bool LIKE_PATTERN>
181
    static VPatternSearchStateSPtr pattern_type_recognition(const ColumnString& patterns);
182
183
    static Status constant_allpass_fn(LikeSearchState* state, const ColumnString& val,
184
                                      const StringRef& pattern, ColumnUInt8::Container& result);
185
186
    static Status constant_allpass_fn_scalar(LikeSearchState* state, const StringRef& val,
187
                                             const StringRef& pattern, unsigned char* result);
188
189
    static Status vector_allpass_fn(const ColumnString& vals, const ColumnString& search_strings,
190
                                    ColumnUInt8::Container& result);
191
192
    static Status constant_starts_with_fn(LikeSearchState* state, const ColumnString& val,
193
                                          const StringRef& pattern, ColumnUInt8::Container& result);
194
195
    static Status constant_starts_with_fn_scalar(LikeSearchState* state, const StringRef& val,
196
                                                 const StringRef& pattern, unsigned char* result);
197
198
    static Status vector_starts_with_fn(const ColumnString& vals,
199
                                        const ColumnString& search_strings,
200
                                        ColumnUInt8::Container& result);
201
202
    static Status constant_ends_with_fn(LikeSearchState* state, const ColumnString& val,
203
                                        const StringRef& pattern, ColumnUInt8::Container& result);
204
205
    static Status constant_ends_with_fn_scalar(LikeSearchState* state, const StringRef& val,
206
                                               const StringRef& pattern, unsigned char* result);
207
208
    static Status vector_ends_with_fn(const ColumnString& vals, const ColumnString& search_strings,
209
                                      ColumnUInt8::Container& result);
210
211
    static Status constant_equals_fn(LikeSearchState* state, const ColumnString& val,
212
                                     const StringRef& pattern, ColumnUInt8::Container& result);
213
214
    static Status constant_equals_fn_scalar(LikeSearchState* state, const StringRef& val,
215
                                            const StringRef& pattern, unsigned char* result);
216
217
    static Status vector_equals_fn(const ColumnString& vals, const ColumnString& search_strings,
218
                                   ColumnUInt8::Container& result);
219
220
    static Status constant_substring_fn(LikeSearchState* state, const ColumnString& val,
221
                                        const StringRef& pattern, ColumnUInt8::Container& result);
222
223
    static Status constant_substring_fn_scalar(LikeSearchState* state, const StringRef& val,
224
                                               const StringRef& pattern, unsigned char* result);
225
226
    static Status vector_substring_fn(const ColumnString& vals, const ColumnString& search_strings,
227
                                      ColumnUInt8::Container& result);
228
229
    static Status constant_regex_fn(LikeSearchState* state, const ColumnString& val,
230
                                    const StringRef& pattern, ColumnUInt8::Container& result);
231
232
    static Status constant_regex_fn_scalar(LikeSearchState* state, const StringRef& val,
233
                                           const StringRef& pattern, unsigned char* result);
234
235
    static Status regexp_fn(LikeSearchState* state, const ColumnString& val,
236
                            const StringRef& pattern, ColumnUInt8::Container& result);
237
238
    static Status regexp_fn_scalar(LikeSearchState* state, const StringRef& val,
239
                                   const StringRef& pattern, unsigned char* result);
240
241
    // hyperscan compile expression to database and allocate scratch space
242
    static Status hs_prepare(FunctionContext* context, const char* expression,
243
                             hs_database_t** database, hs_scratch_t** scratch);
244
};
245
246
class FunctionLike : public FunctionLikeBase {
247
public:
248
    static constexpr auto name = "like";
249
250
5
    static FunctionPtr create() { return std::make_shared<FunctionLike>(); }
251
252
1
    String get_name() const override { return name; }
253
254
    Status open(FunctionContext* context, FunctionContext::FunctionStateScope scope) override;
255
256
    static Status construct_like_const_state(FunctionContext* ctx, const StringRef& pattern,
257
                                             std::shared_ptr<LikeState>& state,
258
                                             bool try_hyperscan = true);
259
260
    friend struct LikeSearchState;
261
    friend struct VectorAllpassSearchState;
262
    friend struct VectorEqualSearchState;
263
    friend struct VectorSubStringSearchState;
264
    friend struct VectorStartsWithSearchState;
265
    friend struct VectorEndsWithSearchState;
266
267
private:
268
    static Status like_fn(LikeSearchState* state, const ColumnString& val, const StringRef& pattern,
269
                          ColumnUInt8::Container& result);
270
271
    static Status like_fn_scalar(LikeSearchState* state, const StringRef& val,
272
                                 const StringRef& pattern, unsigned char* result);
273
274
    static void convert_like_pattern(LikeSearchState* state, const std::string& pattern,
275
                                     std::string* re_pattern);
276
277
    static void remove_escape_character(std::string* search_string);
278
};
279
280
class FunctionRegexpLike : public FunctionLikeBase {
281
public:
282
    static constexpr auto name = "regexp";
283
    static constexpr auto alias = "rlike";
284
285
19
    static FunctionPtr create() { return std::make_shared<FunctionRegexpLike>(); }
286
287
1
    String get_name() const override { return name; }
288
289
    Status open(FunctionContext* context, FunctionContext::FunctionStateScope scope) override;
290
};
291
292
} // namespace doris::vectorized