be/src/exprs/table_function/vexplode_json_object.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/vexplode_json_object.h" |
19 | | |
20 | | #include <glog/logging.h> |
21 | | |
22 | | #include <ostream> |
23 | | |
24 | | #include "common/status.h" |
25 | | #include "core/block/block.h" |
26 | | #include "core/block/column_with_type_and_name.h" |
27 | | #include "core/column/column.h" |
28 | | #include "core/column/column_struct.h" |
29 | | #include "core/string_ref.h" |
30 | | #include "exprs/vexpr.h" |
31 | | #include "exprs/vexpr_context.h" |
32 | | #include "util/jsonb_document.h" |
33 | | #include "util/jsonb_writer.h" |
34 | | |
35 | | namespace doris { |
36 | | #include "common/compile_check_begin.h" |
37 | | |
38 | 0 | VExplodeJsonObjectTableFunction::VExplodeJsonObjectTableFunction() { |
39 | 0 | _fn_name = "vexplode_json_object"; |
40 | 0 | } |
41 | | |
42 | 0 | Status VExplodeJsonObjectTableFunction::process_init(Block* block, RuntimeState* state) { |
43 | 0 | CHECK(_expr_context->root()->children().size() == 1) |
44 | 0 | << "VExplodeJsonObjectTableFunction only support 1 child but has " |
45 | 0 | << _expr_context->root()->children().size(); |
46 | |
|
47 | 0 | RETURN_IF_ERROR(_expr_context->root()->children()[0]->execute_column( |
48 | 0 | _expr_context.get(), block, nullptr, block->rows(), _json_object_column)); |
49 | | |
50 | 0 | return Status::OK(); |
51 | 0 | } |
52 | | |
53 | 0 | void VExplodeJsonObjectTableFunction::process_row(size_t row_idx) { |
54 | 0 | TableFunction::process_row(row_idx); |
55 | |
|
56 | 0 | StringRef text = _json_object_column->get_data_at(row_idx); |
57 | 0 | if (text.data != nullptr) { |
58 | 0 | const JsonbDocument* doc = nullptr; |
59 | 0 | auto st = JsonbDocument::checkAndCreateDocument(text.data, text.size, &doc); |
60 | 0 | if (!st.ok() || !doc || !doc->getValue()) [[unlikely]] { |
61 | | // error jsonb, put null into output, cur_size = 0 , we will insert_default |
62 | 0 | return; |
63 | 0 | } |
64 | | // value is NOT necessary to be deleted since JsonbValue will not allocate memory |
65 | 0 | const JsonbValue* value = doc->getValue(); |
66 | 0 | auto writer = std::make_unique<JsonbWriter>(); |
67 | 0 | if (value->isObject()) { |
68 | 0 | _cur_size = value->numElements(); |
69 | 0 | const auto* obj = value->unpack<ObjectVal>(); |
70 | 0 | _object_pairs.first = |
71 | 0 | ColumnNullable::create(ColumnString::create(), ColumnUInt8::create()); |
72 | 0 | _object_pairs.second = |
73 | 0 | ColumnNullable::create(ColumnString::create(), ColumnUInt8::create()); |
74 | 0 | _object_pairs.first->reserve(_cur_size); |
75 | 0 | _object_pairs.second->reserve(_cur_size); |
76 | 0 | for (const auto& it : *obj) { |
77 | 0 | _object_pairs.first->insert_data(it.getKeyStr(), it.klen()); |
78 | 0 | writer->reset(); |
79 | 0 | writer->writeValue(it.value()); |
80 | 0 | if (it.value()->isNull()) { |
81 | 0 | _object_pairs.second->insert_default(); |
82 | 0 | } else { |
83 | 0 | const std::string_view& jsonb_value = std::string_view( |
84 | 0 | writer->getOutput()->getBuffer(), writer->getOutput()->getSize()); |
85 | 0 | _object_pairs.second->insert_data(jsonb_value.data(), jsonb_value.size()); |
86 | 0 | } |
87 | 0 | } |
88 | 0 | } |
89 | | // we do not support other json type except object |
90 | 0 | } |
91 | 0 | } |
92 | | |
93 | 0 | void VExplodeJsonObjectTableFunction::process_close() { |
94 | 0 | _json_object_column = nullptr; |
95 | 0 | _object_pairs.first = nullptr; |
96 | 0 | _object_pairs.second = nullptr; |
97 | 0 | } |
98 | | |
99 | 0 | void VExplodeJsonObjectTableFunction::get_same_many_values(MutableColumnPtr& column, int length) { |
100 | | // if current is empty map row, also append a default value |
101 | 0 | if (current_empty()) { |
102 | 0 | column->insert_many_defaults(length); |
103 | 0 | return; |
104 | 0 | } |
105 | 0 | ColumnStruct* ret = nullptr; |
106 | | // this _is_nullable is whole output column's nullable |
107 | 0 | if (_is_nullable) { |
108 | | // make map kv value into struct |
109 | 0 | ret = assert_cast<ColumnStruct*>( |
110 | 0 | assert_cast<ColumnNullable*>(column.get())->get_nested_column_ptr().get()); |
111 | 0 | assert_cast<ColumnUInt8*>( |
112 | 0 | assert_cast<ColumnNullable*>(column.get())->get_null_map_column_ptr().get()) |
113 | 0 | ->insert_many_defaults(length); |
114 | 0 | } else if (is_column<ColumnStruct>(column.get())) { |
115 | 0 | ret = assert_cast<ColumnStruct*>(column.get()); |
116 | 0 | } else { |
117 | 0 | throw Exception(ErrorCode::INTERNAL_ERROR, |
118 | 0 | "only support expand json object int to struct(kv pair), but given: ", |
119 | 0 | column->get_name()); |
120 | 0 | } |
121 | 0 | if (!ret || ret->tuple_size() != 2) { |
122 | 0 | throw Exception(ErrorCode::INTERNAL_ERROR, |
123 | 0 | "only support expand json object int to kv pair column, but given: ", |
124 | 0 | ret->tuple_size()); |
125 | 0 | } |
126 | 0 | ret->get_column(0).insert_many_from(*_object_pairs.first, _cur_offset, length); |
127 | 0 | ret->get_column(1).insert_many_from(*_object_pairs.second, _cur_offset, length); |
128 | 0 | } |
129 | | |
130 | 0 | int VExplodeJsonObjectTableFunction::get_value(MutableColumnPtr& column, int max_step) { |
131 | 0 | max_step = std::min(max_step, (int)(_cur_size - _cur_offset)); |
132 | 0 | if (current_empty()) { |
133 | 0 | column->insert_default(); |
134 | 0 | max_step = 1; |
135 | 0 | } else { |
136 | 0 | ColumnStruct* struct_column = nullptr; |
137 | 0 | if (_is_nullable) { |
138 | 0 | auto* nullable_column = assert_cast<ColumnNullable*>(column.get()); |
139 | 0 | struct_column = |
140 | 0 | assert_cast<ColumnStruct*>(nullable_column->get_nested_column_ptr().get()); |
141 | 0 | auto* nullmap_column = |
142 | 0 | assert_cast<ColumnUInt8*>(nullable_column->get_null_map_column_ptr().get()); |
143 | | // here nullmap_column insert max_step many defaults as if MAP[row_idx] is NULL |
144 | | // will be not update value, _cur_size = 0, means current_empty; |
145 | | // so here could insert directly |
146 | 0 | nullmap_column->insert_many_defaults(max_step); |
147 | 0 | } else { |
148 | 0 | struct_column = assert_cast<ColumnStruct*>(column.get()); |
149 | 0 | } |
150 | 0 | if (!struct_column || struct_column->tuple_size() != 2) { |
151 | 0 | throw Exception(ErrorCode::INTERNAL_ERROR, |
152 | 0 | "only support expand json object int to kv pair column, but given: ", |
153 | 0 | struct_column->tuple_size()); |
154 | 0 | } |
155 | 0 | struct_column->get_column(0).insert_range_from(*_object_pairs.first, _cur_offset, max_step); |
156 | 0 | struct_column->get_column(1).insert_range_from(*_object_pairs.second, _cur_offset, |
157 | 0 | max_step); |
158 | 0 | } |
159 | 0 | forward(max_step); |
160 | 0 | return max_step; |
161 | 0 | } |
162 | | |
163 | | #include "common/compile_check_end.h" |
164 | | } // namespace doris |