Coverage Report

Created: 2026-04-15 12:36

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
be/src/exprs/table_function/vjson_each.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 "exprs/table_function/vjson_each.h"
19
20
#include <glog/logging.h>
21
22
#include <ostream>
23
#include <string>
24
25
#include "common/status.h"
26
#include "core/assert_cast.h"
27
#include "core/block/block.h"
28
#include "core/block/column_with_type_and_name.h"
29
#include "core/column/column.h"
30
#include "core/column/column_const.h"
31
#include "core/column/column_struct.h"
32
#include "core/string_ref.h"
33
#include "exprs/vexpr.h"
34
#include "exprs/vexpr_context.h"
35
#include "util/jsonb_document.h"
36
#include "util/jsonb_utils.h"
37
#include "util/jsonb_writer.h"
38
39
namespace doris {
40
41
template <bool TEXT_MODE>
42
23
VJsonEachTableFunction<TEXT_MODE>::VJsonEachTableFunction() {
43
23
    _fn_name = TEXT_MODE ? "vjson_each_text" : "vjson_each";
44
23
}
_ZN5doris22VJsonEachTableFunctionILb0EEC2Ev
Line
Count
Source
42
17
VJsonEachTableFunction<TEXT_MODE>::VJsonEachTableFunction() {
43
17
    _fn_name = TEXT_MODE ? "vjson_each_text" : "vjson_each";
44
17
}
_ZN5doris22VJsonEachTableFunctionILb1EEC2Ev
Line
Count
Source
42
6
VJsonEachTableFunction<TEXT_MODE>::VJsonEachTableFunction() {
43
6
    _fn_name = TEXT_MODE ? "vjson_each_text" : "vjson_each";
44
6
}
45
46
template <bool TEXT_MODE>
47
39
Status VJsonEachTableFunction<TEXT_MODE>::process_init(Block* block, RuntimeState* /*state*/) {
48
39
    ColumnPtr value_column;
49
39
    RETURN_IF_ERROR(_expr_context->root()->children()[0]->execute_column(
50
39
            _expr_context.get(), block, nullptr, block->rows(), value_column));
51
39
    auto [col, is_const] = unpack_if_const(value_column);
52
39
    _json_column = col;
53
39
    _is_const = is_const;
54
39
    return Status::OK();
55
39
}
_ZN5doris22VJsonEachTableFunctionILb0EE12process_initEPNS_5BlockEPNS_12RuntimeStateE
Line
Count
Source
47
26
Status VJsonEachTableFunction<TEXT_MODE>::process_init(Block* block, RuntimeState* /*state*/) {
48
26
    ColumnPtr value_column;
49
26
    RETURN_IF_ERROR(_expr_context->root()->children()[0]->execute_column(
50
26
            _expr_context.get(), block, nullptr, block->rows(), value_column));
51
26
    auto [col, is_const] = unpack_if_const(value_column);
52
26
    _json_column = col;
53
26
    _is_const = is_const;
54
26
    return Status::OK();
55
26
}
_ZN5doris22VJsonEachTableFunctionILb1EE12process_initEPNS_5BlockEPNS_12RuntimeStateE
Line
Count
Source
47
13
Status VJsonEachTableFunction<TEXT_MODE>::process_init(Block* block, RuntimeState* /*state*/) {
48
13
    ColumnPtr value_column;
49
13
    RETURN_IF_ERROR(_expr_context->root()->children()[0]->execute_column(
50
13
            _expr_context.get(), block, nullptr, block->rows(), value_column));
51
13
    auto [col, is_const] = unpack_if_const(value_column);
52
13
    _json_column = col;
53
13
    _is_const = is_const;
54
13
    return Status::OK();
55
13
}
56
57
// Helper: insert one JsonbValue as plain text into a ColumnNullable<ColumnString>.
58
// For strings: raw blob content (quotes stripped, matching json_each_text PG semantics).
59
// For null JSON values: SQL NULL (insert_default).
60
// For all others (numbers, bools, objects, arrays): JSON text representation.
61
12
static void insert_value_as_text(const JsonbValue* value, MutableColumnPtr& col) {
62
12
    if (value == nullptr || value->isNull()) {
63
1
        col->insert_default();
64
1
        return;
65
1
    }
66
11
    if (value->isString()) {
67
6
        const auto* str_val = value->unpack<JsonbStringVal>();
68
6
        col->insert_data(str_val->getBlob(), str_val->getBlobLen());
69
6
    } else {
70
5
        JsonbToJson converter;
71
5
        std::string text = converter.to_json_string(value);
72
5
        col->insert_data(text.data(), text.size());
73
5
    }
74
11
}
75
76
// Helper: insert one JsonbValue in JSONB binary form into a ColumnNullable<ColumnString>.
77
// For null JSON values: SQL NULL (insert_default).
78
// For all others: write JSONB binary via JsonbWriter.
79
static void insert_value_as_json(const JsonbValue* value, MutableColumnPtr& col,
80
26
                                 JsonbWriter& writer) {
81
26
    if (value == nullptr || value->isNull()) {
82
2
        col->insert_default();
83
2
        return;
84
2
    }
85
24
    writer.reset();
86
24
    writer.writeValue(value);
87
24
    const auto* buf = writer.getOutput()->getBuffer();
88
24
    size_t len = writer.getOutput()->getSize();
89
24
    col->insert_data(buf, len);
90
24
}
91
92
template <bool TEXT_MODE>
93
41
void VJsonEachTableFunction<TEXT_MODE>::process_row(size_t row_idx) {
94
41
    TableFunction::process_row(row_idx);
95
41
    if (_is_const && _cur_size > 0) {
96
1
        return;
97
1
    }
98
99
40
    StringRef text;
100
40
    const size_t idx = _is_const ? 0 : row_idx;
101
40
    if (const auto* nullable_col = check_and_get_column<ColumnNullable>(*_json_column)) {
102
40
        if (nullable_col->is_null_at(idx)) {
103
5
            return;
104
5
        }
105
35
        text = assert_cast<const ColumnString&>(nullable_col->get_nested_column()).get_data_at(idx);
106
35
    } else {
107
0
        text = assert_cast<const ColumnString&>(*_json_column).get_data_at(idx);
108
0
    }
109
110
35
    const JsonbDocument* doc = nullptr;
111
35
    auto st = JsonbDocument::checkAndCreateDocument(text.data, text.size, &doc);
112
35
    if (!st.ok() || !doc || !doc->getValue()) [[unlikely]] {
113
3
        return;
114
3
    }
115
116
32
    const JsonbValue* jv = doc->getValue();
117
32
    if (!jv->isObject()) {
118
4
        return;
119
4
    }
120
121
28
    const auto* obj = jv->unpack<ObjectVal>();
122
28
    _cur_size = obj->numElem();
123
28
    if (_cur_size == 0) {
124
7
        return;
125
7
    }
126
127
21
    _kv_pairs.first = ColumnNullable::create(ColumnString::create(), ColumnUInt8::create());
128
21
    _kv_pairs.second = ColumnNullable::create(ColumnString::create(), ColumnUInt8::create());
129
21
    _kv_pairs.first->reserve(_cur_size);
130
21
    _kv_pairs.second->reserve(_cur_size);
131
132
21
    if constexpr (TEXT_MODE) {
133
12
        for (const auto& kv : *obj) {
134
12
            _kv_pairs.first->insert_data(kv.getKeyStr(), kv.klen());
135
12
            insert_value_as_text(kv.value(), _kv_pairs.second);
136
12
        }
137
15
    } else {
138
15
        JsonbWriter writer;
139
26
        for (const auto& kv : *obj) {
140
26
            _kv_pairs.first->insert_data(kv.getKeyStr(), kv.klen());
141
26
            insert_value_as_json(kv.value(), _kv_pairs.second, writer);
142
26
        }
143
15
    }
144
21
}
_ZN5doris22VJsonEachTableFunctionILb0EE11process_rowEm
Line
Count
Source
93
28
void VJsonEachTableFunction<TEXT_MODE>::process_row(size_t row_idx) {
94
28
    TableFunction::process_row(row_idx);
95
28
    if (_is_const && _cur_size > 0) {
96
1
        return;
97
1
    }
98
99
27
    StringRef text;
100
27
    const size_t idx = _is_const ? 0 : row_idx;
101
27
    if (const auto* nullable_col = check_and_get_column<ColumnNullable>(*_json_column)) {
102
27
        if (nullable_col->is_null_at(idx)) {
103
3
            return;
104
3
        }
105
24
        text = assert_cast<const ColumnString&>(nullable_col->get_nested_column()).get_data_at(idx);
106
24
    } else {
107
0
        text = assert_cast<const ColumnString&>(*_json_column).get_data_at(idx);
108
0
    }
109
110
24
    const JsonbDocument* doc = nullptr;
111
24
    auto st = JsonbDocument::checkAndCreateDocument(text.data, text.size, &doc);
112
24
    if (!st.ok() || !doc || !doc->getValue()) [[unlikely]] {
113
1
        return;
114
1
    }
115
116
23
    const JsonbValue* jv = doc->getValue();
117
23
    if (!jv->isObject()) {
118
3
        return;
119
3
    }
120
121
20
    const auto* obj = jv->unpack<ObjectVal>();
122
20
    _cur_size = obj->numElem();
123
20
    if (_cur_size == 0) {
124
5
        return;
125
5
    }
126
127
15
    _kv_pairs.first = ColumnNullable::create(ColumnString::create(), ColumnUInt8::create());
128
15
    _kv_pairs.second = ColumnNullable::create(ColumnString::create(), ColumnUInt8::create());
129
15
    _kv_pairs.first->reserve(_cur_size);
130
15
    _kv_pairs.second->reserve(_cur_size);
131
132
    if constexpr (TEXT_MODE) {
133
        for (const auto& kv : *obj) {
134
            _kv_pairs.first->insert_data(kv.getKeyStr(), kv.klen());
135
            insert_value_as_text(kv.value(), _kv_pairs.second);
136
        }
137
15
    } else {
138
15
        JsonbWriter writer;
139
26
        for (const auto& kv : *obj) {
140
26
            _kv_pairs.first->insert_data(kv.getKeyStr(), kv.klen());
141
26
            insert_value_as_json(kv.value(), _kv_pairs.second, writer);
142
26
        }
143
15
    }
144
15
}
_ZN5doris22VJsonEachTableFunctionILb1EE11process_rowEm
Line
Count
Source
93
13
void VJsonEachTableFunction<TEXT_MODE>::process_row(size_t row_idx) {
94
13
    TableFunction::process_row(row_idx);
95
13
    if (_is_const && _cur_size > 0) {
96
0
        return;
97
0
    }
98
99
13
    StringRef text;
100
13
    const size_t idx = _is_const ? 0 : row_idx;
101
13
    if (const auto* nullable_col = check_and_get_column<ColumnNullable>(*_json_column)) {
102
13
        if (nullable_col->is_null_at(idx)) {
103
2
            return;
104
2
        }
105
11
        text = assert_cast<const ColumnString&>(nullable_col->get_nested_column()).get_data_at(idx);
106
11
    } else {
107
0
        text = assert_cast<const ColumnString&>(*_json_column).get_data_at(idx);
108
0
    }
109
110
11
    const JsonbDocument* doc = nullptr;
111
11
    auto st = JsonbDocument::checkAndCreateDocument(text.data, text.size, &doc);
112
11
    if (!st.ok() || !doc || !doc->getValue()) [[unlikely]] {
113
2
        return;
114
2
    }
115
116
9
    const JsonbValue* jv = doc->getValue();
117
9
    if (!jv->isObject()) {
118
1
        return;
119
1
    }
120
121
8
    const auto* obj = jv->unpack<ObjectVal>();
122
8
    _cur_size = obj->numElem();
123
8
    if (_cur_size == 0) {
124
2
        return;
125
2
    }
126
127
6
    _kv_pairs.first = ColumnNullable::create(ColumnString::create(), ColumnUInt8::create());
128
6
    _kv_pairs.second = ColumnNullable::create(ColumnString::create(), ColumnUInt8::create());
129
6
    _kv_pairs.first->reserve(_cur_size);
130
6
    _kv_pairs.second->reserve(_cur_size);
131
132
6
    if constexpr (TEXT_MODE) {
133
12
        for (const auto& kv : *obj) {
134
12
            _kv_pairs.first->insert_data(kv.getKeyStr(), kv.klen());
135
12
            insert_value_as_text(kv.value(), _kv_pairs.second);
136
12
        }
137
    } else {
138
        JsonbWriter writer;
139
        for (const auto& kv : *obj) {
140
            _kv_pairs.first->insert_data(kv.getKeyStr(), kv.klen());
141
            insert_value_as_json(kv.value(), _kv_pairs.second, writer);
142
        }
143
    }
144
6
}
145
146
template <bool TEXT_MODE>
147
39
void VJsonEachTableFunction<TEXT_MODE>::process_close() {
148
39
    _json_column = nullptr;
149
39
    _kv_pairs.first = nullptr;
150
39
    _kv_pairs.second = nullptr;
151
39
    _cur_size = 0;
152
39
}
_ZN5doris22VJsonEachTableFunctionILb0EE13process_closeEv
Line
Count
Source
147
26
void VJsonEachTableFunction<TEXT_MODE>::process_close() {
148
26
    _json_column = nullptr;
149
26
    _kv_pairs.first = nullptr;
150
26
    _kv_pairs.second = nullptr;
151
26
    _cur_size = 0;
152
26
}
_ZN5doris22VJsonEachTableFunctionILb1EE13process_closeEv
Line
Count
Source
147
13
void VJsonEachTableFunction<TEXT_MODE>::process_close() {
148
13
    _json_column = nullptr;
149
13
    _kv_pairs.first = nullptr;
150
13
    _kv_pairs.second = nullptr;
151
13
    _cur_size = 0;
152
13
}
153
154
template <bool TEXT_MODE>
155
5
void VJsonEachTableFunction<TEXT_MODE>::get_same_many_values(MutableColumnPtr& column, int length) {
156
5
    if (current_empty()) {
157
1
        column->insert_many_defaults(length);
158
1
        return;
159
1
    }
160
161
4
    ColumnStruct* ret;
162
4
    if (_is_nullable) {
163
3
        auto* nullable = assert_cast<ColumnNullable*>(column.get());
164
3
        ret = assert_cast<ColumnStruct*>(nullable->get_nested_column_ptr().get());
165
3
        assert_cast<ColumnUInt8*>(nullable->get_null_map_column_ptr().get())
166
3
                ->insert_many_defaults(length);
167
3
    } else {
168
1
        ret = assert_cast<ColumnStruct*>(column.get());
169
1
    }
170
171
4
    ret->get_column(0).insert_many_from(*_kv_pairs.first, _cur_offset, length);
172
4
    ret->get_column(1).insert_many_from(*_kv_pairs.second, _cur_offset, length);
173
4
}
_ZN5doris22VJsonEachTableFunctionILb0EE20get_same_many_valuesERNS_3COWINS_7IColumnEE11mutable_ptrIS3_EEi
Line
Count
Source
155
4
void VJsonEachTableFunction<TEXT_MODE>::get_same_many_values(MutableColumnPtr& column, int length) {
156
4
    if (current_empty()) {
157
1
        column->insert_many_defaults(length);
158
1
        return;
159
1
    }
160
161
3
    ColumnStruct* ret;
162
3
    if (_is_nullable) {
163
2
        auto* nullable = assert_cast<ColumnNullable*>(column.get());
164
2
        ret = assert_cast<ColumnStruct*>(nullable->get_nested_column_ptr().get());
165
2
        assert_cast<ColumnUInt8*>(nullable->get_null_map_column_ptr().get())
166
2
                ->insert_many_defaults(length);
167
2
    } else {
168
1
        ret = assert_cast<ColumnStruct*>(column.get());
169
1
    }
170
171
3
    ret->get_column(0).insert_many_from(*_kv_pairs.first, _cur_offset, length);
172
3
    ret->get_column(1).insert_many_from(*_kv_pairs.second, _cur_offset, length);
173
3
}
_ZN5doris22VJsonEachTableFunctionILb1EE20get_same_many_valuesERNS_3COWINS_7IColumnEE11mutable_ptrIS3_EEi
Line
Count
Source
155
1
void VJsonEachTableFunction<TEXT_MODE>::get_same_many_values(MutableColumnPtr& column, int length) {
156
1
    if (current_empty()) {
157
0
        column->insert_many_defaults(length);
158
0
        return;
159
0
    }
160
161
1
    ColumnStruct* ret;
162
1
    if (_is_nullable) {
163
1
        auto* nullable = assert_cast<ColumnNullable*>(column.get());
164
1
        ret = assert_cast<ColumnStruct*>(nullable->get_nested_column_ptr().get());
165
1
        assert_cast<ColumnUInt8*>(nullable->get_null_map_column_ptr().get())
166
1
                ->insert_many_defaults(length);
167
1
    } else {
168
0
        ret = assert_cast<ColumnStruct*>(column.get());
169
0
    }
170
171
1
    ret->get_column(0).insert_many_from(*_kv_pairs.first, _cur_offset, length);
172
1
    ret->get_column(1).insert_many_from(*_kv_pairs.second, _cur_offset, length);
173
1
}
174
175
template <bool TEXT_MODE>
176
28
int VJsonEachTableFunction<TEXT_MODE>::get_value(MutableColumnPtr& column, int max_step) {
177
28
    max_step = std::min(max_step, (int)(_cur_size - _cur_offset));
178
179
28
    if (current_empty()) {
180
9
        column->insert_default();
181
9
        max_step = 1;
182
19
    } else {
183
19
        ColumnStruct* struct_col = nullptr;
184
19
        if (_is_nullable) {
185
18
            auto* nullable_col = assert_cast<ColumnNullable*>(column.get());
186
18
            struct_col = assert_cast<ColumnStruct*>(nullable_col->get_nested_column_ptr().get());
187
18
            assert_cast<ColumnUInt8*>(nullable_col->get_null_map_column_ptr().get())
188
18
                    ->insert_many_defaults(max_step);
189
18
        } else {
190
1
            struct_col = assert_cast<ColumnStruct*>(column.get());
191
1
        }
192
193
19
        struct_col->get_column(0).insert_range_from(*_kv_pairs.first, _cur_offset, max_step);
194
19
        struct_col->get_column(1).insert_range_from(*_kv_pairs.second, _cur_offset, max_step);
195
19
    }
196
197
28
    forward(max_step);
198
28
    return max_step;
199
28
}
_ZN5doris22VJsonEachTableFunctionILb0EE9get_valueERNS_3COWINS_7IColumnEE11mutable_ptrIS3_EEi
Line
Count
Source
176
16
int VJsonEachTableFunction<TEXT_MODE>::get_value(MutableColumnPtr& column, int max_step) {
177
16
    max_step = std::min(max_step, (int)(_cur_size - _cur_offset));
178
179
16
    if (current_empty()) {
180
5
        column->insert_default();
181
5
        max_step = 1;
182
11
    } else {
183
11
        ColumnStruct* struct_col = nullptr;
184
11
        if (_is_nullable) {
185
10
            auto* nullable_col = assert_cast<ColumnNullable*>(column.get());
186
10
            struct_col = assert_cast<ColumnStruct*>(nullable_col->get_nested_column_ptr().get());
187
10
            assert_cast<ColumnUInt8*>(nullable_col->get_null_map_column_ptr().get())
188
10
                    ->insert_many_defaults(max_step);
189
10
        } else {
190
1
            struct_col = assert_cast<ColumnStruct*>(column.get());
191
1
        }
192
193
11
        struct_col->get_column(0).insert_range_from(*_kv_pairs.first, _cur_offset, max_step);
194
11
        struct_col->get_column(1).insert_range_from(*_kv_pairs.second, _cur_offset, max_step);
195
11
    }
196
197
16
    forward(max_step);
198
16
    return max_step;
199
16
}
_ZN5doris22VJsonEachTableFunctionILb1EE9get_valueERNS_3COWINS_7IColumnEE11mutable_ptrIS3_EEi
Line
Count
Source
176
12
int VJsonEachTableFunction<TEXT_MODE>::get_value(MutableColumnPtr& column, int max_step) {
177
12
    max_step = std::min(max_step, (int)(_cur_size - _cur_offset));
178
179
12
    if (current_empty()) {
180
4
        column->insert_default();
181
4
        max_step = 1;
182
8
    } else {
183
8
        ColumnStruct* struct_col = nullptr;
184
8
        if (_is_nullable) {
185
8
            auto* nullable_col = assert_cast<ColumnNullable*>(column.get());
186
8
            struct_col = assert_cast<ColumnStruct*>(nullable_col->get_nested_column_ptr().get());
187
8
            assert_cast<ColumnUInt8*>(nullable_col->get_null_map_column_ptr().get())
188
8
                    ->insert_many_defaults(max_step);
189
8
        } else {
190
0
            struct_col = assert_cast<ColumnStruct*>(column.get());
191
0
        }
192
193
8
        struct_col->get_column(0).insert_range_from(*_kv_pairs.first, _cur_offset, max_step);
194
8
        struct_col->get_column(1).insert_range_from(*_kv_pairs.second, _cur_offset, max_step);
195
8
    }
196
197
12
    forward(max_step);
198
12
    return max_step;
199
12
}
200
201
// // Explicit template instantiations
202
template class VJsonEachTableFunction<false>; // json_each
203
template class VJsonEachTableFunction<true>;  // json_each_text
204
205
} // namespace doris