org.apache.druid.sql

Class PreparedStatement

java.lang.Object

org.apache.druid.sql.AbstractStatement

org.apache.druid.sql.PreparedStatement

All Implemented Interfaces:

Closeable, AutoCloseable

public class PreparedStatement
extends AbstractStatement

Statement for the JDBC prepare-once, execute many model.

Field Summary

Fields inherited from class org.apache.druid.sql.AbstractStatement

cancellationResourceActions, fullResourceActions, plannerContext, queryPlus, reporter, sqlToolbox

Constructor Summary

Constructors

Constructor and Description
PreparedStatement(SqlToolbox lifecycleToolbox, SqlQueryPlus queryPlus)

Method Summary

Modifier and Type	Method and Description
DirectStatement	execute(List<org.apache.calcite.avatica.remote.TypedValue> parameters) Execute a prepared JDBC query.
PrepareResult	prepare() Prepare the query lifecycle for execution, without completely planning into something that is executable, but including some initial parsing and validation, to support prepared statements via JDBC.

Methods inherited from class org.apache.druid.sql.AbstractStatement

allResources, authorize, authorizer, close, closeQuietly, closeWithError, query, reporter, resources, sqlQueryId, validate

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

PreparedStatement

public PreparedStatement(SqlToolbox lifecycleToolbox,
                         SqlQueryPlus queryPlus)

Method Detail

prepare

public PrepareResult prepare()

Prepare the query lifecycle for execution, without completely planning into something that is executable, but including some initial parsing and validation, to support prepared statements via JDBC.

Note that, per JDBC convention, the prepare step does not provide parameter values: those are provided later during execution and will generally vary from one execution to the next.

• Create the planner.
• Parse the statement.
• JDBC does not provide parameter values at prepare time. They are provided during execution later, where we'll replan the query to use the "query optimized" structure.
• Validate the query against the Druid catalog.
• Authorize access to the resources which the query needs.
• Return a PrepareResult which describes the query.

execute

public DirectStatement execute(List<org.apache.calcite.avatica.remote.TypedValue> parameters)

Execute a prepared JDBC query. Druid uses "query optimized" parameters, which means we do not reuse the statement prepared above, but rather plan anew with the actual parameter values. The same statement can be execute many times, including concurrently. Each execution repeats the parse, validate, authorize and plan steps since data, permissions, views and other dependencies may have changed.