Coverage Report

Created: 2025-04-24 16:21

/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/columns_number.h"
39
#include "vec/columns/predicate_column.h"
40
#include "vec/common/string_ref.h"
41
#include "vec/core/column_numbers.h"
42
#include "vec/core/types.h"
43
#include "vec/data_types/data_type_number.h"
44
#include "vec/functions/function.h"
45
46
namespace doris {
47
namespace vectorized {
48
class Block;
49
} // namespace vectorized
50
} // namespace doris
51
52
namespace doris::vectorized {
53
54
// TODO: replace with std::string_view when `LikeSearchState.substring_pattern` can
55
// construct from std::string_view.
56
struct LikeSearchState {
57
    char escape_char;
58
59
    /// Holds the string the StringRef points to and is set any time StringRef is
60
    /// used.
61
    std::string search_string;
62
63
    std::string pattern_str;
64
65
    /// Used for LIKE predicates if the pattern is a constant argument, and is either a
66
    /// constant string or has a constant string at the beginning or end of the pattern.
67
    /// This will be set in order to check for that pattern in the corresponding part of
68
    /// the string.
69
    StringRef search_string_sv;
70
71
    /// Used for LIKE predicates if the pattern is a constant argument and has a constant
72
    /// string in the middle of it. This will be use in order to check for the substring
73
    /// in the value.
74
    doris::StringSearch substring_pattern;
75
76
    /// Used for RLIKE and REGEXP predicates if the pattern is a constant argument.
77
    std::unique_ptr<re2::RE2> regex;
78
79
    template <typename Deleter, Deleter deleter>
80
    struct HyperscanDeleter {
81
        template <typename T>
82
10
        void operator()(T* ptr) const {
83
10
            deleter(ptr);
84
10
        }
_ZNK5doris10vectorized15LikeSearchState16HyperscanDeleterIPFiP11hs_databaseEXadL_Z16hs_free_databaseEEEclIS3_EEvPT_
Line
Count
Source
82
5
        void operator()(T* ptr) const {
83
5
            deleter(ptr);
84
5
        }
_ZNK5doris10vectorized15LikeSearchState16HyperscanDeleterIPFiP10hs_scratchEXadL_Z15hs_free_scratchEEEclIS3_EEvPT_
Line
Count
Source
82
5
        void operator()(T* ptr) const {
83
5
            deleter(ptr);
84
5
        }
85
    };
86
87
    // hyperscan compiled pattern database and scratch space, reused for performance
88
    std::unique_ptr<hs_database_t, HyperscanDeleter<decltype(&hs_free_database), &hs_free_database>>
89
            hs_database;
90
    std::unique_ptr<hs_scratch_t, HyperscanDeleter<decltype(&hs_free_scratch), &hs_free_scratch>>
91
            hs_scratch;
92
93
    // hyperscan match callback
94
    static int hs_match_handler(unsigned int /* from */,       // NOLINT
95
                                unsigned long long /* from */, // NOLINT
96
                                unsigned long long /* to */,   // NOLINT
97
3
                                unsigned int /* flags */, void* ctx) {
98
        // set result to 1 for matched row
99
3
        *((unsigned char*)ctx) = 1;
100
        /// return non-zero to indicate hyperscan stop after first matched
101
3
        return 1;
102
3
    }
103
104
18
    LikeSearchState() : escape_char('\\') {}
105
106
    Status clone(LikeSearchState& cloned);
107
108
12
    void set_search_string(const std::string& search_string_arg) {
109
12
        search_string = search_string_arg;
110
12
        search_string_sv = StringRef(search_string);
111
12
        substring_pattern.set_pattern(&search_string_sv);
112
12
    }
113
};
114
115
using LikeFn = std::function<doris::Status(LikeSearchState*, const ColumnString&, const StringRef&,
116
                                           ColumnUInt8::Container&)>;
117
118
using ScalarLikeFn = std::function<doris::Status(LikeSearchState*, const StringRef&,
119
                                                 const StringRef&, unsigned char*)>;
120
121
using VectorLikeFn = std::function<doris::Status(const ColumnString&, const ColumnString&,
122
                                                 ColumnUInt8::Container&)>;
123
124
struct LikeState {
125
    bool is_like_pattern;
126
    LikeSearchState search_state;
127
    LikeFn function;
128
    ScalarLikeFn scalar_function;
129
};
130
131
struct VectorPatternSearchState {
132
    MutableColumnPtr _search_strings;
133
    std::string _search_string;
134
    VectorLikeFn _vector_function;
135
    bool _pattern_matched;
136
137
    VectorPatternSearchState(VectorLikeFn vector_function)
138
            : _search_strings(ColumnString::create()),
139
              _vector_function(vector_function),
140
60
              _pattern_matched(true) {}
141
142
60
    virtual ~VectorPatternSearchState() = default;
143
144
    virtual void like_pattern_match(const std::string& pattern_str) = 0;
145
146
    virtual void regexp_pattern_match(const std::string& pattern_str) = 0;
147
};
148
149
using VPatternSearchStateSPtr = std::shared_ptr<VectorPatternSearchState>;
150
151
class FunctionLikeBase : public IFunction {
152
public:
153
20
    size_t get_number_of_arguments() const override { return 2; }
154
155
20
    DataTypePtr get_return_type_impl(const DataTypes& /*arguments*/) const override {
156
20
        return std::make_shared<DataTypeUInt8>();
157
20
    }
158
159
    Status execute_impl(FunctionContext* context, Block& block, const ColumnNumbers& arguments,
160
                        uint32_t result, size_t /*input_rows_count*/) const override;
161
162
    friend struct VectorAllpassSearchState;
163
    friend struct VectorEqualSearchState;
164
    friend struct VectorSubStringSearchState;
165
    friend struct VectorStartsWithSearchState;
166
    friend struct VectorEndsWithSearchState;
167
168
protected:
169
    Status vector_const(const ColumnString& values, const StringRef* pattern_val,
170
                        ColumnUInt8::Container& result, const LikeFn& function,
171
                        LikeSearchState* search_state) const;
172
173
    Status vector_non_const(const ColumnString& values, const ColumnString& patterns,
174
                            ColumnUInt8::Container& result, LikeState* state,
175
                            size_t input_rows_count) const;
176
177
    Status execute_substring(const ColumnString::Chars& values,
178
                             const ColumnString::Offsets& value_offsets,
179
                             ColumnUInt8::Container& result, LikeSearchState* search_state) const;
180
181
    template <bool LIKE_PATTERN>
182
    static VPatternSearchStateSPtr pattern_type_recognition(const ColumnString& patterns);
183
184
    static Status constant_allpass_fn(LikeSearchState* state, const ColumnString& val,
185
                                      const StringRef& pattern, ColumnUInt8::Container& result);
186
187
    static Status constant_allpass_fn_scalar(LikeSearchState* state, const StringRef& val,
188
                                             const StringRef& pattern, unsigned char* result);
189
190
    static Status vector_allpass_fn(const ColumnString& vals, const ColumnString& search_strings,
191
                                    ColumnUInt8::Container& result);
192
193
    static Status constant_starts_with_fn(LikeSearchState* state, const ColumnString& val,
194
                                          const StringRef& pattern, ColumnUInt8::Container& result);
195
196
    static Status constant_starts_with_fn_scalar(LikeSearchState* state, const StringRef& val,
197
                                                 const StringRef& pattern, unsigned char* result);
198
199
    static Status vector_starts_with_fn(const ColumnString& vals,
200
                                        const ColumnString& search_strings,
201
                                        ColumnUInt8::Container& result);
202
203
    static Status constant_ends_with_fn(LikeSearchState* state, const ColumnString& val,
204
                                        const StringRef& pattern, ColumnUInt8::Container& result);
205
206
    static Status constant_ends_with_fn_scalar(LikeSearchState* state, const StringRef& val,
207
                                               const StringRef& pattern, unsigned char* result);
208
209
    static Status vector_ends_with_fn(const ColumnString& vals, const ColumnString& search_strings,
210
                                      ColumnUInt8::Container& result);
211
212
    static Status constant_equals_fn(LikeSearchState* state, const ColumnString& val,
213
                                     const StringRef& pattern, ColumnUInt8::Container& result);
214
215
    static Status constant_equals_fn_scalar(LikeSearchState* state, const StringRef& val,
216
                                            const StringRef& pattern, unsigned char* result);
217
218
    static Status vector_equals_fn(const ColumnString& vals, const ColumnString& search_strings,
219
                                   ColumnUInt8::Container& result);
220
221
    static Status constant_substring_fn(LikeSearchState* state, const ColumnString& val,
222
                                        const StringRef& pattern, ColumnUInt8::Container& result);
223
224
    static Status constant_substring_fn_scalar(LikeSearchState* state, const StringRef& val,
225
                                               const StringRef& pattern, unsigned char* result);
226
227
    static Status vector_substring_fn(const ColumnString& vals, const ColumnString& search_strings,
228
                                      ColumnUInt8::Container& result);
229
230
    static Status constant_regex_fn(LikeSearchState* state, const ColumnString& val,
231
                                    const StringRef& pattern, ColumnUInt8::Container& result);
232
233
    static Status constant_regex_fn_scalar(LikeSearchState* state, const StringRef& val,
234
                                           const StringRef& pattern, unsigned char* result);
235
236
    static Status regexp_fn(LikeSearchState* state, const ColumnString& val,
237
                            const StringRef& pattern, ColumnUInt8::Container& result);
238
239
    static Status regexp_fn_scalar(LikeSearchState* state, const StringRef& val,
240
                                   const StringRef& pattern, unsigned char* result);
241
242
    // hyperscan compile expression to database and allocate scratch space
243
    static Status hs_prepare(FunctionContext* context, const char* expression,
244
                             hs_database_t** database, hs_scratch_t** scratch);
245
};
246
247
class FunctionLike : public FunctionLikeBase {
248
public:
249
    static constexpr auto name = "like";
250
251
5
    static FunctionPtr create() { return std::make_shared<FunctionLike>(); }
252
253
1
    String get_name() const override { return name; }
254
255
    Status open(FunctionContext* context, FunctionContext::FunctionStateScope scope) override;
256
257
    static Status construct_like_const_state(FunctionContext* ctx, const StringRef& pattern,
258
                                             std::shared_ptr<LikeState>& state,
259
                                             bool try_hyperscan = true);
260
261
    friend struct LikeSearchState;
262
    friend struct VectorAllpassSearchState;
263
    friend struct VectorEqualSearchState;
264
    friend struct VectorSubStringSearchState;
265
    friend struct VectorStartsWithSearchState;
266
    friend struct VectorEndsWithSearchState;
267
268
private:
269
    static Status like_fn(LikeSearchState* state, const ColumnString& val, const StringRef& pattern,
270
                          ColumnUInt8::Container& result);
271
272
    static Status like_fn_scalar(LikeSearchState* state, const StringRef& val,
273
                                 const StringRef& pattern, unsigned char* result);
274
275
    static void convert_like_pattern(LikeSearchState* state, const std::string& pattern,
276
                                     std::string* re_pattern);
277
278
    static void remove_escape_character(std::string* search_string);
279
};
280
281
class FunctionRegexpLike : public FunctionLikeBase {
282
public:
283
    static constexpr auto name = "regexp";
284
    static constexpr auto alias = "rlike";
285
286
19
    static FunctionPtr create() { return std::make_shared<FunctionRegexpLike>(); }
287
288
1
    String get_name() const override { return name; }
289
290
    Status open(FunctionContext* context, FunctionContext::FunctionStateScope scope) override;
291
};
292
293
} // namespace doris::vectorized