/Users/deen/code/yugabyte-db/src/yb/yql/pggate/pg_dml.h
Line | Count | Source |
1 | | //-------------------------------------------------------------------------------------------------- |
2 | | // Copyright (c) YugaByte, Inc. |
3 | | // |
4 | | // Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except |
5 | | // in compliance with the License. You may obtain a copy of the License at |
6 | | // |
7 | | // http://www.apache.org/licenses/LICENSE-2.0 |
8 | | // |
9 | | // Unless required by applicable law or agreed to in writing, software distributed under the License |
10 | | // is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express |
11 | | // or implied. See the License for the specific language governing permissions and limitations |
12 | | // under the License. |
13 | | //-------------------------------------------------------------------------------------------------- |
14 | | |
15 | | #ifndef YB_YQL_PGGATE_PG_DML_H_ |
16 | | #define YB_YQL_PGGATE_PG_DML_H_ |
17 | | |
18 | | #include <boost/unordered_map.hpp> |
19 | | |
20 | | #include "yb/yql/pggate/pg_doc_op.h" |
21 | | #include "yb/yql/pggate/pg_session.h" |
22 | | #include "yb/yql/pggate/pg_statement.h" |
23 | | #include "yb/yql/pggate/pg_table.h" |
24 | | |
25 | | namespace yb { |
26 | | namespace pggate { |
27 | | |
28 | | //-------------------------------------------------------------------------------------------------- |
29 | | // DML |
30 | | //-------------------------------------------------------------------------------------------------- |
31 | | class PgSelectIndex; |
32 | | |
33 | | class PgDml : public PgStatement { |
34 | | public: |
35 | | virtual ~PgDml(); |
36 | | |
37 | | // Append a target in SELECT or RETURNING. |
38 | | CHECKED_STATUS AppendTarget(PgExpr *target); |
39 | | |
40 | | // Append a filter condition. |
41 | | // Supported expression kind is serialized Postgres expression |
42 | | CHECKED_STATUS AppendQual(PgExpr *qual); |
43 | | |
44 | | // Append a column reference. |
45 | | // If any serialized Postgres expressions appended to other lists require explicit addition |
46 | | // of their column references. Those column references should have Postgres type information. |
47 | | // Other PgExpr kinds are automatically scanned and their column references are appended. |
48 | | CHECKED_STATUS AppendColumnRef(PgExpr *colref); |
49 | | |
50 | | // Prepare column for both ends. |
51 | | // - Prepare protobuf to communicate with DocDB. |
52 | | // - Prepare PgExpr to send data back to Postgres layer. |
53 | | Result<const PgColumn&> PrepareColumnForRead(int attr_num, PgsqlExpressionPB *target_pb); |
54 | | CHECKED_STATUS PrepareColumnForWrite(PgColumn *pg_col, PgsqlExpressionPB *assign_pb); |
55 | | |
56 | | // Bind a column with an expression. |
57 | | // - For a secondary-index-scan, this bind specify the value of the secondary key which is used to |
58 | | // query a row. |
59 | | // - For a primary-index-scan, this bind specify the value of the keys of the table. |
60 | | CHECKED_STATUS BindColumn(int attnum, PgExpr *attr_value); |
61 | | |
62 | | // Bind the whole table. |
63 | | CHECKED_STATUS BindTable(); |
64 | | |
65 | | // Assign an expression to a column. |
66 | | CHECKED_STATUS AssignColumn(int attnum, PgExpr *attr_value); |
67 | | |
68 | | // Process the secondary index request if it is nested within this statement. |
69 | | Result<bool> ProcessSecondaryIndexRequest(const PgExecParameters *exec_params); |
70 | | |
71 | | // Fetch a row and return it to Postgres layer. |
72 | | CHECKED_STATUS Fetch(int32_t natts, |
73 | | uint64_t *values, |
74 | | bool *isnulls, |
75 | | PgSysColumns *syscols, |
76 | | bool *has_data); |
77 | | |
78 | | // Returns TRUE if docdb replies with more data. |
79 | | Result<bool> FetchDataFromServer(); |
80 | | |
81 | | // Returns TRUE if desired row is found. |
82 | | Result<bool> GetNextRow(PgTuple *pg_tuple); |
83 | | |
84 | | virtual void SetCatalogCacheVersion(uint64_t catalog_cache_version) = 0; |
85 | | |
86 | | // Get column info on whether the column 'attr_num' is a hash key, a range |
87 | | // key, or neither. |
88 | | Result<YBCPgColumnInfo> GetColumnInfo(int attr_num) const; |
89 | | |
90 | | bool has_aggregate_targets(); |
91 | | |
92 | 5.39M | bool has_doc_op() const { |
93 | 5.39M | return doc_op_ != nullptr; |
94 | 5.39M | } |
95 | | |
96 | | protected: |
97 | | // Method members. |
98 | | // Constructor. |
99 | | PgDml(PgSession::ScopedRefPtr pg_session, const PgObjectId& table_id); |
100 | | PgDml(PgSession::ScopedRefPtr pg_session, |
101 | | const PgObjectId& table_id, |
102 | | const PgObjectId& index_id, |
103 | | const PgPrepareParameters *prepare_params); |
104 | | |
105 | | // Allocate protobuf for a SELECTed expression. |
106 | | virtual PgsqlExpressionPB *AllocTargetPB() = 0; |
107 | | |
108 | | // Allocate protobuf for a WHERE clause expression. |
109 | | // Subclasses use different protobuf message types for their requests, so they should |
110 | | // implement this method that knows how to add a PgsqlExpressionPB entry into their where_clauses |
111 | | // field. |
112 | | virtual PgsqlExpressionPB *AllocQualPB() = 0; |
113 | | |
114 | | // Allocate protobuf for expression whose value is bounded to a column. |
115 | | virtual PgsqlExpressionPB *AllocColumnBindPB(PgColumn *col) = 0; |
116 | | |
117 | | // Allocate protobuf for expression whose value is assigned to a column (SET clause). |
118 | | virtual PgsqlExpressionPB *AllocColumnAssignPB(PgColumn *col) = 0; |
119 | | |
120 | | // Specify target of the query in protobuf request. |
121 | | CHECKED_STATUS AppendTargetPB(PgExpr *target); |
122 | | |
123 | | // Update bind values. |
124 | | CHECKED_STATUS UpdateBindPBs(); |
125 | | |
126 | | // Update set values. |
127 | | CHECKED_STATUS UpdateAssignPBs(); |
128 | | |
129 | | // Compatibility: set deprecated column_refs for legacy nodes |
130 | | // We are deprecating PgsqlColumnRefsPB protobuf since it does not allow to transfer Postgres |
131 | | // type information required to evaluate serialized Postgres expressions. |
132 | | // It is being replaced by list of PgsqlColRefPB entries, which is set by ColRefsToPB. |
133 | | // While there is are chance of cluster being upgraded from older version, we have to populate |
134 | | // both. |
135 | | void ColumnRefsToPB(PgsqlColumnRefsPB *column_refs); |
136 | | |
137 | | // Transfer columns information from target_.columns() to the request's col_refs list field. |
138 | | // Subclasses use different protobuf message types to make requests, so they must implement |
139 | | // the ClearColRefPBs and AllocColRefPB virtual methods to respectively remove all old col_refs |
140 | | // entries and allocate new entry in their requests. |
141 | | void ColRefsToPB(); |
142 | | |
143 | | // Clear previously allocated PgsqlColRefPB entries from the protobuf request |
144 | | virtual void ClearColRefPBs() = 0; |
145 | | |
146 | | // Allocate a PgsqlColRefPB entriy in the protobuf request |
147 | | virtual PgsqlColRefPB *AllocColRefPB() = 0; |
148 | | |
149 | | // ----------------------------------------------------------------------------------------------- |
150 | | // Data members that define the DML statement. |
151 | | |
152 | | // Table identifiers |
153 | | // - table_id_ identifies the table to read data from. |
154 | | // - index_id_ identifies the index to be used for scanning. |
155 | | // |
156 | | // Example for query on table_id_ using index_id_. |
157 | | // SELECT FROM "table_id_" |
158 | | // WHERE ybctid IN (SELECT base_ybctid FROM "index_id_" WHERE matched-index-binds) |
159 | | // |
160 | | // - Postgres will create PgSelect(table_id_) { nested PgSelectIndex (index_id_) } |
161 | | // - When bind functions are called, it bind user-values to columns in PgSelectIndex as these |
162 | | // binds will be used to find base_ybctid from the IndexTable. |
163 | | // - When AddTargets() is called, the target is added to PgSlect as data will be reading from |
164 | | // table_id_ using the found base_ybctid from index_id_. |
165 | | PgObjectId table_id_; |
166 | | PgObjectId index_id_; |
167 | | |
168 | | // Targets of statements (Output parameter). |
169 | | // - "target_desc_" is the table descriptor where data will be read from. |
170 | | // - "targets_" are either selected or returned expressions by DML statements. |
171 | | PgTable target_; |
172 | | std::vector<PgExpr*> targets_; |
173 | | |
174 | | // Qual is a where clause condition pushed to the DocDB to filter scanned rows |
175 | | // Qual supports PgExprs holding serialized Postgres expressions, and require the column |
176 | | // references used in these Quals to be explicitly added with AppendColumnRef() |
177 | | std::vector<PgExpr*> quals_; |
178 | | |
179 | | // bind_desc_ is the descriptor of the table whose key columns' values will be specified by the |
180 | | // the DML statement being executed. |
181 | | // - For primary key binding, "bind_desc_" is the descriptor of the main table as we don't have |
182 | | // a separated primary-index table. |
183 | | // - For secondary key binding, "bind_desc_" is the descriptor of teh secondary index table. |
184 | | // The bound values will be used to read base_ybctid which is then used to read actual data |
185 | | // from the main table. |
186 | | PgTable bind_; |
187 | | |
188 | | // Prepare control parameters. |
189 | | PgPrepareParameters prepare_params_ = { .index_oid = kInvalidOid, |
190 | | .index_only_scan = false, |
191 | | .use_secondary_index = false, |
192 | | .querying_colocated_table = false }; |
193 | | |
194 | | // ----------------------------------------------------------------------------------------------- |
195 | | // Data members for nested query: This is used for an optimization in PgGate. |
196 | | // |
197 | | // - Each DML operation can be understood as |
198 | | // Read / Write TABLE WHERE ybctid IN (SELECT ybctid from INDEX). |
199 | | // - In most cases, the Postgres layer processes the subquery "SELECT ybctid from INDEX". |
200 | | // - Under certain conditions, to optimize the performance, the PgGate layer might operate on |
201 | | // the INDEX subquery itself. |
202 | | std::unique_ptr<PgSelectIndex> secondary_index_query_; |
203 | | |
204 | | // ----------------------------------------------------------------------------------------------- |
205 | | // Data members for generated protobuf. |
206 | | // NOTE: |
207 | | // - Where clause processing data is not supported yet. |
208 | | // - Some protobuf structure are also set up in PgColumn class. |
209 | | |
210 | | // Column associated values (expressions) to be used by DML statements. |
211 | | // - When expression are constructed, we bind them with their associated protobuf. |
212 | | // - These expressions might not yet have values for place_holders or literals. |
213 | | // - During execution, the place_holder values are updated, and the statement protobuf need to |
214 | | // be updated accordingly. |
215 | | // |
216 | | // * Bind values are used to identify the selected rows to be operated on. |
217 | | // * Set values are used to hold columns' new values in the selected rows. |
218 | | bool ybctid_bind_ = false; |
219 | | boost::unordered_map<PgsqlExpressionPB*, PgExpr*> expr_binds_; |
220 | | std::unordered_map<PgsqlExpressionPB*, PgExpr*> expr_assigns_; |
221 | | |
222 | | // Used for colocated TRUNCATE that doesn't bind any columns. |
223 | | bool bind_table_ = false; |
224 | | |
225 | | // DML Operator. |
226 | | PgDocOp::SharedPtr doc_op_; |
227 | | |
228 | | //------------------------------------------------------------------------------------------------ |
229 | | // Data members for navigating the output / result-set from either seleted or returned targets. |
230 | | std::list<PgDocResult> rowsets_; |
231 | | int64_t current_row_order_ = 0; |
232 | | |
233 | | // Yugabyte has a few IN/OUT parameters of statement execution, "pg_exec_params_" is used to sent |
234 | | // OUT value back to postgres. |
235 | | const PgExecParameters *pg_exec_params_ = NULL; |
236 | | |
237 | | //------------------------------------------------------------------------------------------------ |
238 | | // Hashed and range values/components used to compute the tuple id. |
239 | | // |
240 | | // These members are populated by the AddYBTupleIdColumn function and the tuple id is retrieved |
241 | | // using the GetYBTupleId function. |
242 | | // |
243 | | // These members are not used internally by the statement and are simply a utility for computing |
244 | | // the tuple id (ybctid). |
245 | | }; |
246 | | |
247 | | } // namespace pggate |
248 | | } // namespace yb |
249 | | |
250 | | #endif // YB_YQL_PGGATE_PG_DML_H_ |