Coverage Report

Created: 2026-05-26 17:00

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
be/src/util/jsonb/serialize.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 "util/jsonb/serialize.h"
19
20
#include <assert.h>
21
22
#include <algorithm>
23
#include <limits>
24
#include <memory>
25
#include <unordered_set>
26
#include <vector>
27
28
#include "core/arena.h"
29
#include "core/block/column_with_type_and_name.h"
30
#include "core/block/columns_with_type_and_name.h"
31
#include "core/column/column.h"
32
#include "core/column/column_string.h"
33
#include "core/data_type/data_type.h"
34
#include "core/data_type/primitive_type.h"
35
#include "core/data_type_serde/data_type_serde.h"
36
#include "core/string_ref.h"
37
#include "core/value/jsonb_value.h"
38
#include "runtime/descriptors.h"
39
#include "storage/tablet/tablet_schema.h"
40
#include "util/defer_op.h"
41
#include "util/jsonb_document.h"
42
#include "util/jsonb_stream.h"
43
#include "util/jsonb_writer.h"
44
45
namespace doris {
46
47
void JsonbSerializeUtil::block_to_jsonb(const TabletSchema& schema, const Block& block,
48
                                        ColumnString& dst, int num_cols,
49
                                        const DataTypeSerDeSPtrs& serdes,
50
277
                                        const std::unordered_set<int32_t>& row_store_cids) {
51
277
    block_to_jsonb(schema, block, dst, num_cols, serdes, row_store_cids, 0, block.rows());
52
277
}
53
54
void JsonbSerializeUtil::block_to_jsonb(const TabletSchema& schema, const Block& block,
55
                                        ColumnString& dst, int num_cols,
56
                                        const DataTypeSerDeSPtrs& serdes,
57
                                        const std::unordered_set<int32_t>& row_store_cids,
58
277
                                        size_t row_pos, size_t num_rows) {
59
277
    static_cast<void>(block_to_jsonb(schema, block, dst, num_cols, serdes, row_store_cids, row_pos,
60
277
                                     num_rows, std::numeric_limits<size_t>::max()));
61
277
}
62
63
size_t JsonbSerializeUtil::block_to_jsonb(const TabletSchema& schema, const Block& block,
64
                                          ColumnString& dst, int num_cols,
65
                                          const DataTypeSerDeSPtrs& serdes,
66
                                          const std::unordered_set<int32_t>& row_store_cids,
67
2.40k
                                          size_t row_pos, size_t num_rows, size_t max_bytes) {
68
2.40k
    Arena arena;
69
2.40k
    assert(num_cols <= block.columns());
70
2.40k
    assert(row_pos + num_rows <= block.rows());
71
2.40k
    DataTypeSerDe::FormatOptions options;
72
2.40k
    auto tz = cctz::utc_time_zone();
73
2.40k
    options.timezone = &tz;
74
2.40k
    size_t written_rows = 0;
75
1.91M
    for (size_t i = row_pos; i < row_pos + num_rows; ++i) {
76
1.91M
        JsonbWriterT<JsonbOutStream> jsonb_writer;
77
1.91M
        jsonb_writer.writeStartObject();
78
39.0M
        for (int j = 0; j < num_cols; ++j) {
79
37.1M
            const auto& column = block.get_by_position(j).column;
80
37.1M
            const auto& tablet_column = *schema.columns()[j];
81
            // ignore row store columns
82
37.1M
            if (tablet_column.is_row_store_column()) {
83
1.91M
                continue;
84
1.91M
            }
85
            // TODO improve performance for checking column in group
86
35.2M
            if (row_store_cids.empty() || row_store_cids.contains(tablet_column.unique_id())) {
87
23.7M
                serdes[j]->write_one_cell_to_jsonb(*column, jsonb_writer, arena,
88
23.7M
                                                   tablet_column.unique_id(), i, options);
89
23.7M
            }
90
35.2M
        }
91
1.91M
        jsonb_writer.writeEndObject();
92
1.91M
        dst.insert_data(jsonb_writer.getOutput()->getBuffer(), jsonb_writer.getOutput()->getSize());
93
1.91M
        ++written_rows;
94
1.91M
        if (dst.byte_size() >= max_bytes) {
95
0
            break;
96
0
        }
97
1.91M
    }
98
2.40k
    return written_rows;
99
2.40k
}
100
101
// batch rows
102
Status JsonbSerializeUtil::jsonb_to_block(
103
        const DataTypeSerDeSPtrs& serdes, const ColumnString& jsonb_column,
104
        const std::unordered_map<uint32_t, uint32_t>& col_id_to_idx, Block& dst,
105
        const std::vector<std::string>& default_values,
106
342
        const std::unordered_set<int>& include_cids) {
107
342
    auto dst_columns_guard = dst.mutate_columns_scoped();
108
342
    MutableColumns& dst_columns = dst_columns_guard.mutable_columns();
109
2.60k
    for (int i = 0; i < jsonb_column.size(); ++i) {
110
2.26k
        StringRef jsonb_data = jsonb_column.get_data_at(i);
111
2.26k
        RETURN_IF_ERROR(jsonb_to_columns(serdes, jsonb_data.data, jsonb_data.size, col_id_to_idx,
112
2.26k
                                         dst_columns, default_values, include_cids));
113
2.26k
    }
114
342
    return Status::OK();
115
342
}
116
117
Status JsonbSerializeUtil::jsonb_to_columns(
118
        const DataTypeSerDeSPtrs& serdes, const char* data, size_t size,
119
        const std::unordered_map<uint32_t, uint32_t>& col_id_to_idx, MutableColumns& dst_columns,
120
        const std::vector<std::string>& default_values,
121
6.89k
        const std::unordered_set<int>& include_cids) {
122
6.89k
    const JsonbDocument* pdoc = nullptr;
123
6.89k
    RETURN_IF_ERROR(JsonbDocument::checkAndCreateDocument(data, size, &pdoc));
124
6.89k
    const JsonbDocument& doc = *pdoc;
125
6.89k
    DCHECK(!dst_columns.empty());
126
6.89k
    size_t num_rows = dst_columns[0]->size();
127
6.89k
    size_t filled_columns = 0;
128
116k
    for (auto it = doc->begin(); it != doc->end(); ++it) {
129
109k
        auto col_it = col_id_to_idx.find(it->getKeyId());
130
109k
        if (col_it != col_id_to_idx.end() &&
131
109k
            (include_cids.empty() || include_cids.contains(it->getKeyId()))) {
132
94.7k
            auto& dst_column = dst_columns[col_it->second];
133
94.7k
            serdes[col_it->second]->read_one_cell_from_jsonb(*dst_column, it->value());
134
94.7k
            ++filled_columns;
135
94.7k
        }
136
109k
    }
137
6.89k
    if (filled_columns >= dst_columns.size()) {
138
5.69k
        return Status::OK();
139
5.69k
    }
140
16.6k
    auto fill_column = [&](size_t pos, size_t old_num_rows) {
141
16.6k
        auto& dst_column = dst_columns[pos];
142
16.6k
        if (dst_column->size() < old_num_rows + 1) {
143
1.22k
            DCHECK(dst_column->size() == old_num_rows);
144
1.22k
            Status st = Status::OK();
145
1.22k
            if (default_values[pos].empty()) {
146
1.18k
                dst_column->insert_default();
147
1.18k
            } else {
148
39
                Slice value(default_values[pos].data(), default_values[pos].size());
149
39
                DataTypeSerDe::FormatOptions opt;
150
39
                opt.converted_from_string = true;
151
39
                st = serdes[pos]->deserialize_one_cell_from_json(*dst_column, value, opt);
152
39
            }
153
1.22k
            RETURN_IF_ERROR(st);
154
1.22k
            DCHECK(dst_column->size() == num_rows + 1);
155
1.22k
            return Status::OK();
156
1.22k
        }
157
16.6k
        DCHECK(dst_column->size() == num_rows + 1);
158
15.4k
        return Status::OK();
159
16.6k
    };
160
    // fill missing column
161
1.20k
    if (!include_cids.empty()) {
162
18
        for (auto cid : include_cids) {
163
18
            auto col_it = col_id_to_idx.find(cid);
164
18
            if (col_it == col_id_to_idx.end()) {
165
2
                continue;
166
2
            }
167
16
            RETURN_IF_ERROR(fill_column(static_cast<size_t>(col_it->second), num_rows));
168
16
        }
169
1.19k
    } else {
170
17.8k
        for (size_t i = 0; i < dst_columns.size(); ++i) {
171
16.6k
            RETURN_IF_ERROR(fill_column(i, num_rows));
172
16.6k
        }
173
1.19k
    }
174
1.20k
    return Status::OK();
175
1.20k
}
176
177
// single row
178
Status JsonbSerializeUtil::jsonb_to_block(
179
        const DataTypeSerDeSPtrs& serdes, const char* data, size_t size,
180
        const std::unordered_map<uint32_t, uint32_t>& col_id_to_idx, Block& dst,
181
        const std::vector<std::string>& default_values,
182
0
        const std::unordered_set<int>& include_cids) {
183
0
    auto dst_columns_guard = dst.mutate_columns_scoped();
184
0
    MutableColumns& dst_columns = dst_columns_guard.mutable_columns();
185
0
    return jsonb_to_columns(serdes, data, size, col_id_to_idx, dst_columns, default_values,
186
0
                            include_cids);
187
0
}
188
189
} // namespace doris