Execution plan
An execution process generated by the optimizer for an SQL statement, which is generally expressed as a tree of operators.
Local plan
If an execution plan involves only a single table or a single partition of a partitioned table, and the table or partition is on the current node, the plan is defined as a local plan.
Remote plan
If an execution plan involves only a single table or a single partition of a partitioned table, and the table or partition is on another node, the plan is defined as a remote plan.
Distributed plan
If an execution plan involves multiple tables or partitions, the plan is defined as a distributed plan.
Access path
The method by which a table is accessed, including primary table access and secondary index access. For details, see Overview.
Join order
The order in which tables are joined in a multi-table join. Currently, OceanBase AI Database supports only left-deep join trees. For details, see Join order.
Join algorithm
The algorithm used to perform a join between two tables, including three types: Nested Loop Join, Merge Join, and Hash Join.
Query rewrite
The process of rewriting a user query into an equivalent form to facilitate the generation of an optimal execution plan.
Execution plan binding
The process by which a user specifies the execution plan for an SQL statement by providing an outline. For details, see Plan binding.
SQL plan management (SPM)
A plan evolution mechanism. When the optimizer generates a new plan, the evolution mechanism ensures that the plan's performance does not regress. If regression occurs, the plan is rejected; otherwise, it is used.
Adaptive cursor sharing (ACS)
A mechanism that allows the optimizer to store multiple plans for each parameterized SQL statement and select an appropriate plan based on the selectivity of predicates in the SQL statement.
Data flow object (DFO)
A distributed plan is divided into logically parallel subplans based on data redistribution points, with each subplan encapsulated by a DFO.