Package io.trino.spi.connector

Interface ConnectorMetadata

public interface ConnectorMetadata

Nested Class Summary

Nested Classes

Modifier and Type	Interface	Description
static final class	ConnectorMetadata.Helper

Field Summary

Fields

Modifier and Type	Field	Description
static final String	MODIFYING_ROWS_MESSAGE

Method Summary

Modifier and Type	Method	Description
default void	addColumn(ConnectorSession session, ConnectorTableHandle tableHandle, ColumnMetadata column)	Add the specified column
default void	addField(ConnectorSession session, ConnectorTableHandle tableHandle, List<String> parentPath, String fieldName, Type type, boolean ignoreExisting)	Add the specified field, potentially nested, to a row.
default Optional<AggregationApplicationResult<ConnectorTableHandle>>	applyAggregation(ConnectorSession session, ConnectorTableHandle handle, List<AggregateFunction> aggregates, Map<String,ColumnHandle> assignments, List<List<ColumnHandle>> groupingSets)	Attempt to push down the aggregates into the table.
default Optional<ConnectorTableHandle>	applyDelete(ConnectorSession session, ConnectorTableHandle handle)	Attempt to push down a delete operation into the connector.
default Optional<ConstraintApplicationResult<ConnectorTableHandle>>	applyFilter(ConnectorSession session, ConnectorTableHandle handle, Constraint constraint)	Attempt to push down the provided constraint into the table.
default Optional<JoinApplicationResult<ConnectorTableHandle>>	applyJoin(ConnectorSession session, JoinType joinType, ConnectorTableHandle left, ConnectorTableHandle right, ConnectorExpression joinCondition, Map<String,ColumnHandle> leftAssignments, Map<String,ColumnHandle> rightAssignments, JoinStatistics statistics)	Attempt to push down the join operation.
default Optional<JoinApplicationResult<ConnectorTableHandle>>	applyJoin(ConnectorSession session, JoinType joinType, ConnectorTableHandle left, ConnectorTableHandle right, List<JoinCondition> joinConditions, Map<String,ColumnHandle> leftAssignments, Map<String,ColumnHandle> rightAssignments, JoinStatistics statistics)	Deprecated.
default Optional<LimitApplicationResult<ConnectorTableHandle>>	applyLimit(ConnectorSession session, ConnectorTableHandle handle, long limit)	Attempt to push down the provided limit into the table.
default Optional<ProjectionApplicationResult<ConnectorTableHandle>>	applyProjection(ConnectorSession session, ConnectorTableHandle handle, List<ConnectorExpression> projections, Map<String,ColumnHandle> assignments)	Attempt to push down the provided projections into the table.
default Optional<SampleApplicationResult<ConnectorTableHandle>>	applySample(ConnectorSession session, ConnectorTableHandle handle, SampleType sampleType, double sampleRatio)	Attempt to push down the sampling into the table.
default Optional<TableFunctionApplicationResult<ConnectorTableHandle>>	applyTableFunction(ConnectorSession session, ConnectorTableFunctionHandle handle)	Attempt to push down the table function invocation into the connector.
default Optional<TableScanRedirectApplicationResult>	applyTableScanRedirect(ConnectorSession session, ConnectorTableHandle tableHandle)
default Optional<TopNApplicationResult<ConnectorTableHandle>>	applyTopN(ConnectorSession session, ConnectorTableHandle handle, long topNCount, List<SortItem> sortItems, Map<String,ColumnHandle> assignments)	Attempt to push down the TopN into the table scan.
default Optional<ConnectorTableHandle>	applyUpdate(ConnectorSession session, ConnectorTableHandle handle, Map<ColumnHandle,Constant> assignments)	Attempt to push down an update operation into the connector.
default ConnectorOutputTableHandle	beginCreateTable(ConnectorSession session, ConnectorTableMetadata tableMetadata, Optional<ConnectorTableLayout> layout, RetryMode retryMode)	Deprecated. use beginCreateTable(ConnectorSession session, ConnectorTableMetadata tableMetadata, Optional layout, RetryMode retryMode, boolean replace)
default ConnectorOutputTableHandle	beginCreateTable(ConnectorSession session, ConnectorTableMetadata tableMetadata, Optional<ConnectorTableLayout> layout, RetryMode retryMode, boolean replace)	Begin the atomic creation of a table with data.
default ConnectorInsertTableHandle	beginInsert(ConnectorSession session, ConnectorTableHandle tableHandle, List<ColumnHandle> columns, RetryMode retryMode)	Begin insert query.
default ConnectorMergeTableHandle	beginMerge(ConnectorSession session, ConnectorTableHandle tableHandle, RetryMode retryMode)	Do whatever is necessary to start an MERGE query, returning the ConnectorMergeTableHandle instance that will be passed to the PageSink, and to the finishMerge(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.ConnectorMergeTableHandle, java.util.Collection<io.airlift.slice.Slice>, java.util.Collection<io.trino.spi.statistics.ComputedStatistics>) method.
default void	beginQuery(ConnectorSession session)	Start a query.
default ConnectorInsertTableHandle	beginRefreshMaterializedView(ConnectorSession session, ConnectorTableHandle tableHandle, List<ConnectorTableHandle> sourceTableHandles, RetryMode retryMode)	Begin materialized view query.
default ConnectorTableHandle	beginStatisticsCollection(ConnectorSession session, ConnectorTableHandle tableHandle)	Begin statistics collection
default BeginTableExecuteResult<ConnectorTableExecuteHandle,ConnectorTableHandle>	beginTableExecute(ConnectorSession session, ConnectorTableExecuteHandle tableExecuteHandle, ConnectorTableHandle updatedSourceTableHandle)	Begin execution of table procedure
default void	cleanupQuery(ConnectorSession session)	Cleanup after a query.
default void	createLanguageFunction(ConnectorSession session, SchemaFunctionName name, LanguageFunction function, boolean replace)	Creates a language function with the specified name and signature token.
default void	createMaterializedView(ConnectorSession session, SchemaTableName viewName, ConnectorMaterializedViewDefinition definition, boolean replace, boolean ignoreExisting)	Create the specified materialized view.
default void	createRole(ConnectorSession session, String role, Optional<TrinoPrincipal> grantor)	Creates the specified role.
default void	createSchema(ConnectorSession session, String schemaName, Map<String,Object> properties, TrinoPrincipal owner)	Creates a schema.
default void	createTable(ConnectorSession session, ConnectorTableMetadata tableMetadata, boolean ignoreExisting)	Deprecated. use createTable(ConnectorSession session, ConnectorTableMetadata tableMetadata, SaveMode saveMode)
default void	createTable(ConnectorSession session, ConnectorTableMetadata tableMetadata, SaveMode saveMode)	Creates a table using the specified table metadata.
default void	createView(ConnectorSession session, SchemaTableName viewName, ConnectorViewDefinition definition, boolean replace)	Create the specified view.
default boolean	delegateMaterializedViewRefreshToConnector(ConnectorSession session, SchemaTableName viewName)	Returns true if materialized view refresh should be delegated to connector using refreshMaterializedView(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.SchemaTableName)
default void	denySchemaPrivileges(ConnectorSession session, String schemaName, Set<Privilege> privileges, TrinoPrincipal grantee)	Denys the specified privilege to the specified user on the specified schema
default void	denyTablePrivileges(ConnectorSession session, SchemaTableName tableName, Set<Privilege> privileges, TrinoPrincipal grantee)	Denys the specified privilege to the specified user on the specified table
default void	dropColumn(ConnectorSession session, ConnectorTableHandle tableHandle, ColumnHandle column)	Drop the specified column
default void	dropField(ConnectorSession session, ConnectorTableHandle tableHandle, ColumnHandle column, List<String> fieldPath)	Drop the specified field, potentially nested, from a row.
default void	dropLanguageFunction(ConnectorSession session, SchemaFunctionName name, String signatureToken)	Drops a language function with the specified name and signature token.
default void	dropMaterializedView(ConnectorSession session, SchemaTableName viewName)	Drop the specified materialized view.
default void	dropRole(ConnectorSession session, String role)	Drops the specified role.
default void	dropSchema(ConnectorSession session, String schemaName, boolean cascade)	Drops the specified schema.
default void	dropTable(ConnectorSession session, ConnectorTableHandle tableHandle)	Drops the specified table
default void	dropView(ConnectorSession session, SchemaTableName viewName)	Drop the specified view.
default OptionalLong	executeDelete(ConnectorSession session, ConnectorTableHandle handle)	Execute the delete operation on the handle returned from applyDelete(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.ConnectorTableHandle).
default void	executeTableExecute(ConnectorSession session, ConnectorTableExecuteHandle tableExecuteHandle)	Execute a coordinator-only table procedure.
default OptionalLong	executeUpdate(ConnectorSession session, ConnectorTableHandle handle)	Execute the update operation on the handle returned from applyUpdate(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.ConnectorTableHandle, java.util.Map<io.trino.spi.connector.ColumnHandle, io.trino.spi.expression.Constant>).
default Optional<ConnectorOutputMetadata>	finishCreateTable(ConnectorSession session, ConnectorOutputTableHandle tableHandle, Collection<io.airlift.slice.Slice> fragments, Collection<ComputedStatistics> computedStatistics)	Finish a table creation with data after the data is written.
default Optional<ConnectorOutputMetadata>	finishInsert(ConnectorSession session, ConnectorInsertTableHandle insertHandle, Collection<io.airlift.slice.Slice> fragments, Collection<ComputedStatistics> computedStatistics)	Finish insert query
default void	finishMerge(ConnectorSession session, ConnectorMergeTableHandle mergeTableHandle, Collection<io.airlift.slice.Slice> fragments, Collection<ComputedStatistics> computedStatistics)	Finish a merge query
default Optional<ConnectorOutputMetadata>	finishRefreshMaterializedView(ConnectorSession session, ConnectorTableHandle tableHandle, ConnectorInsertTableHandle insertHandle, Collection<io.airlift.slice.Slice> fragments, Collection<ComputedStatistics> computedStatistics, List<ConnectorTableHandle> sourceTableHandles)	Finish materialized view query
default void	finishStatisticsCollection(ConnectorSession session, ConnectorTableHandle tableHandle, Collection<ComputedStatistics> computedStatistics)	Finish statistics collection
default void	finishTableExecute(ConnectorSession session, ConnectorTableExecuteHandle tableExecuteHandle, Collection<io.airlift.slice.Slice> fragments, List<Object> tableExecuteState)	Finish table execute
default AggregationFunctionMetadata	getAggregationFunctionMetadata(ConnectorSession session, FunctionId functionId)	Returns the aggregation metadata for the aggregation function with the specified id.
default Map<String,ColumnHandle>	getColumnHandles(ConnectorSession session, ConnectorTableHandle tableHandle)	Gets all of the columns on the specified table, or an empty map if the columns cannot be enumerated.
default ColumnMetadata	getColumnMetadata(ConnectorSession session, ConnectorTableHandle tableHandle, ColumnHandle columnHandle)	Gets the metadata for the specified table column.
default Optional<ConnectorPartitioningHandle>	getCommonPartitioningHandle(ConnectorSession session, ConnectorPartitioningHandle left, ConnectorPartitioningHandle right)	Return a partitioning handle which the connector can transparently convert both left and right into.
default FunctionDependencyDeclaration	getFunctionDependencies(ConnectorSession session, FunctionId functionId, BoundSignature boundSignature)	Returns the dependencies of the function with the specified id.
default FunctionMetadata	getFunctionMetadata(ConnectorSession session, FunctionId functionId)	Return the function with the specified id.
default Collection<FunctionMetadata>	getFunctions(ConnectorSession session, SchemaFunctionName name)	Get all functions with specified name.
default Optional<Object>	getInfo(ConnectorTableHandle table)
default Optional<ConnectorTableLayout>	getInsertLayout(ConnectorSession session, ConnectorTableHandle tableHandle)	Get the physical layout for inserting into an existing table.
default WriterScalingOptions	getInsertWriterScalingOptions(ConnectorSession session, ConnectorTableHandle tableHandle)
default Collection<LanguageFunction>	getLanguageFunctions(ConnectorSession session, SchemaFunctionName name)	Get all language functions with the specified name.
default Optional<ConnectorTableLayout>	getLayoutForTableExecute(ConnectorSession session, ConnectorTableExecuteHandle tableExecuteHandle)
default Optional<ConnectorMaterializedViewDefinition>	getMaterializedView(ConnectorSession session, SchemaTableName viewName)	Gets the materialized view data for the specified materialized view name.
default MaterializedViewFreshness	getMaterializedViewFreshness(ConnectorSession session, SchemaTableName name)	The method is used by the engine to determine if a materialized view is current with respect to the tables it depends on.
default Map<SchemaTableName,ConnectorMaterializedViewDefinition>	getMaterializedViews(ConnectorSession session, Optional<String> schemaName)	Gets the definitions of materialized views, possibly filtered by schema.
default OptionalInt	getMaxWriterTasks(ConnectorSession session)
default ColumnHandle	getMergeRowIdColumnHandle(ConnectorSession session, ConnectorTableHandle tableHandle)	Get the column handle that will generate row IDs for the merge operation.
default Optional<ConnectorTableLayout>	getNewTableLayout(ConnectorSession session, ConnectorTableMetadata tableMetadata)	Get the physical layout for a new table.
default WriterScalingOptions	getNewTableWriterScalingOptions(ConnectorSession session, SchemaTableName tableName, Map<String,Object> tableProperties)
default RowChangeParadigm	getRowChangeParadigm(ConnectorSession session, ConnectorTableHandle tableHandle)	Return the row change paradigm supported by the connector on the table.
default Optional<TrinoPrincipal>	getSchemaOwner(ConnectorSession session, String schemaName)	Get the schema properties for the specified schema.
default Map<String,Object>	getSchemaProperties(ConnectorSession session, String schemaName)	Gets the schema properties for the specified schema.
default ConnectorAnalyzeMetadata	getStatisticsCollectionMetadata(ConnectorSession session, ConnectorTableHandle tableHandle, Map<String,Object> analyzeProperties)	Describe statistics that must be collected during a statistics collection
default TableStatisticsMetadata	getStatisticsCollectionMetadataForWrite(ConnectorSession session, ConnectorTableMetadata tableMetadata)	Describes statistics that must be collected during a write.
default Optional<Type>	getSupportedType(ConnectorSession session, Map<String,Object> tableProperties, Type type)	Return the effective Type that is supported by the connector for the given type.
default Optional<SystemTable>	getSystemTable(ConnectorSession session, SchemaTableName tableName)	Returns the system table for the specified table name, if one exists.
default ConnectorTableHandle	getTableHandle(ConnectorSession session, SchemaTableName tableName)	Deprecated. Implement getTableHandle(ConnectorSession, SchemaTableName, Optional, Optional).
default ConnectorTableHandle	getTableHandle(ConnectorSession session, SchemaTableName tableName, Optional<ConnectorTableVersion> startVersion, Optional<ConnectorTableVersion> endVersion)	Returns a table handle for the specified table name and version, or null if tableName relation does not exist or is not a table (e.g.
default Optional<ConnectorTableExecuteHandle>	getTableHandleForExecute(ConnectorSession session, ConnectorTableHandle tableHandle, String procedureName, Map<String,Object> executeProperties, RetryMode retryMode)	Create initial handle for execution of table procedure.
default ConnectorTableMetadata	getTableMetadata(ConnectorSession session, ConnectorTableHandle table)	Return the metadata for the specified table handle.
default SchemaTableName	getTableName(ConnectorSession session, ConnectorTableHandle table)	Return schema table name for the specified table handle.
default ConnectorTableProperties	getTableProperties(ConnectorSession session, ConnectorTableHandle table)
default ConnectorTableSchema	getTableSchema(ConnectorSession session, ConnectorTableHandle table)	Return table schema definition for the specified table handle.
default TableStatistics	getTableStatistics(ConnectorSession session, ConnectorTableHandle tableHandle)	Get statistics for table.
default Optional<ConnectorPartitioningHandle>	getUpdateLayout(ConnectorSession session, ConnectorTableHandle tableHandle)	Get the physical layout for updated or deleted rows of a MERGE operation.
default Optional<ConnectorViewDefinition>	getView(ConnectorSession session, SchemaTableName viewName)	Gets the view data for the specified view name.
default Map<SchemaTableName,ConnectorViewDefinition>	getViews(ConnectorSession session, Optional<String> schemaName)	Gets the definitions of views (but not materialized views), possibly filtered by schema.
default void	grantRoles(ConnectorSession connectorSession, Set<String> roles, Set<TrinoPrincipal> grantees, boolean adminOption, Optional<TrinoPrincipal> grantor)	Grants the specified roles to the specified grantees
default void	grantSchemaPrivileges(ConnectorSession session, String schemaName, Set<Privilege> privileges, TrinoPrincipal grantee, boolean grantOption)	Grants the specified privilege to the specified user on the specified schema
default void	grantTablePrivileges(ConnectorSession session, SchemaTableName tableName, Set<Privilege> privileges, TrinoPrincipal grantee, boolean grantOption)	Grants the specified privilege to the specified user on the specified table
default boolean	languageFunctionExists(ConnectorSession session, SchemaFunctionName name, String signatureToken)	Check if a language function exists.
default Set<RoleGrant>	listApplicableRoles(ConnectorSession session, TrinoPrincipal principal)	List applicable roles, including the transitive grants, for the specified principal
default Set<String>	listEnabledRoles(ConnectorSession session)	List applicable roles, including the transitive grants, in given session
default Collection<FunctionMetadata>	listFunctions(ConnectorSession session, String schemaName)	List available functions.
default Collection<LanguageFunction>	listLanguageFunctions(ConnectorSession session, String schemaName)	List available language functions.
default List<SchemaTableName>	listMaterializedViews(ConnectorSession session, Optional<String> schemaName)	List materialized view names, possibly filtered by schema.
default Set<RoleGrant>	listRoleGrants(ConnectorSession session, TrinoPrincipal principal)	List role grants for a given principal, not recursively.
default Set<String>	listRoles(ConnectorSession session)	List available roles.
default List<String>	listSchemaNames(ConnectorSession session)	Returns the schemas provided by this connector.
default Map<SchemaTableName,List<ColumnMetadata>>	listTableColumns(ConnectorSession session, SchemaTablePrefix prefix)	Deprecated. use streamTableColumns(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.SchemaTablePrefix) which handles redirected tables
default List<GrantInfo>	listTablePrivileges(ConnectorSession session, SchemaTablePrefix prefix)	List the table privileges granted to the specified grantee for the tables that have the specified prefix considering the selected session role
default List<SchemaTableName>	listTables(ConnectorSession session, Optional<String> schemaName)	List table, view and materialized view names, possibly filtered by schema.
default List<SchemaTableName>	listViews(ConnectorSession session, Optional<String> schemaName)	List view names (but not materialized views), possibly filtered by schema.
default ConnectorTableHandle	makeCompatiblePartitioning(ConnectorSession session, ConnectorTableHandle tableHandle, ConnectorPartitioningHandle partitioningHandle)	Return a table handle whose partitioning is converted to the provided partitioning handle, but otherwise identical to the provided table handle.
default Optional<CatalogSchemaTableName>	redirectTable(ConnectorSession session, SchemaTableName tableName)	Redirects table to other table which may or may not be in the same catalog.
default CompletableFuture<?>	refreshMaterializedView(ConnectorSession session, SchemaTableName viewName)	Refresh materialized view
default void	renameColumn(ConnectorSession session, ConnectorTableHandle tableHandle, ColumnHandle source, String target)	Rename the specified column
default void	renameField(ConnectorSession session, ConnectorTableHandle tableHandle, List<String> fieldPath, String target)	Rename the specified field, potentially nested, to a row.
default void	renameMaterializedView(ConnectorSession session, SchemaTableName source, SchemaTableName target)	Rename the specified materialized view
default void	renameSchema(ConnectorSession session, String source, String target)	Renames the specified schema.
default void	renameTable(ConnectorSession session, ConnectorTableHandle tableHandle, SchemaTableName newTableName)	Rename the specified table
default void	renameView(ConnectorSession session, SchemaTableName source, SchemaTableName target)	Rename the specified view
default Optional<ConnectorResolvedIndex>	resolveIndex(ConnectorSession session, ConnectorTableHandle tableHandle, Set<ColumnHandle> indexableColumns, Set<ColumnHandle> outputColumns, TupleDomain<ColumnHandle> tupleDomain)	Try to locate a table index that can lookup results by indexableColumns and provide the requested outputColumns.
default void	revokeRoles(ConnectorSession connectorSession, Set<String> roles, Set<TrinoPrincipal> grantees, boolean adminOption, Optional<TrinoPrincipal> grantor)	Revokes the specified roles from the specified grantees
default void	revokeSchemaPrivileges(ConnectorSession session, String schemaName, Set<Privilege> privileges, TrinoPrincipal grantee, boolean grantOption)	Revokes the specified privilege on the specified schema from the specified user
default void	revokeTablePrivileges(ConnectorSession session, SchemaTableName tableName, Set<Privilege> privileges, TrinoPrincipal grantee, boolean grantOption)	Revokes the specified privilege on the specified table from the specified user
default boolean	roleExists(ConnectorSession session, String role)	Does the specified role exist.
default boolean	schemaExists(ConnectorSession session, String schemaName)	Checks if a schema exists.
default void	setColumnComment(ConnectorSession session, ConnectorTableHandle tableHandle, ColumnHandle column, Optional<String> comment)	Comments to the specified column
default void	setColumnType(ConnectorSession session, ConnectorTableHandle tableHandle, ColumnHandle column, Type type)	Set the specified column type
default void	setFieldType(ConnectorSession session, ConnectorTableHandle tableHandle, List<String> fieldPath, Type type)	Set the specified field type
default void	setMaterializedViewColumnComment(ConnectorSession session, SchemaTableName viewName, String columnName, Optional<String> comment)	Comments to the specified materialized view column.
default void	setMaterializedViewProperties(ConnectorSession session, SchemaTableName viewName, Map<String,Optional<Object>> properties)	Sets the properties of the specified materialized view
default void	setSchemaAuthorization(ConnectorSession session, String schemaName, TrinoPrincipal principal)	Sets the user/role on the specified schema.
default void	setTableAuthorization(ConnectorSession session, SchemaTableName tableName, TrinoPrincipal principal)	Sets the user/role on the specified table.
default void	setTableComment(ConnectorSession session, ConnectorTableHandle tableHandle, Optional<String> comment)	Comments to the specified table
default void	setTableProperties(ConnectorSession session, ConnectorTableHandle tableHandle, Map<String,Optional<Object>> properties)	Set properties to the specified table
default void	setViewAuthorization(ConnectorSession session, SchemaTableName viewName, TrinoPrincipal principal)	Sets the user/role on the specified view.
default void	setViewColumnComment(ConnectorSession session, SchemaTableName viewName, String columnName, Optional<String> comment)	Comments to the specified view column.
default void	setViewComment(ConnectorSession session, SchemaTableName viewName, Optional<String> comment)	Comments to the specified view
default Iterator<RelationColumnsMetadata>	streamRelationColumns(ConnectorSession session, Optional<String> schemaName, UnaryOperator<Set<SchemaTableName>> relationFilter)	Gets columns for all relations (tables, views, materialized views), possibly filtered by schemaName.
default Iterator<RelationCommentMetadata>	streamRelationComments(ConnectorSession session, Optional<String> schemaName, UnaryOperator<Set<SchemaTableName>> relationFilter)	Gets comments for all relations (tables, views, materialized views), possibly filtered by schemaName.
default Iterator<TableColumnsMetadata>	streamTableColumns(ConnectorSession session, SchemaTablePrefix prefix)	Deprecated. Implement streamRelationColumns(io.trino.spi.connector.ConnectorSession, java.util.Optional<java.lang.String>, java.util.function.UnaryOperator<java.util.Set<io.trino.spi.connector.SchemaTableName>>).
default boolean	supportsMissingColumnsOnInsert()
default void	truncateTable(ConnectorSession session, ConnectorTableHandle tableHandle)	Truncates the specified table
default void	validateScan(ConnectorSession session, ConnectorTableHandle handle)	Allows the connector to reject the table scan produced by the planner.

Field Details

MODIFYING_ROWS_MESSAGE

static final String MODIFYING_ROWS_MESSAGE

See Also:

• Constant Field Values

Method Details

schemaExists

default boolean schemaExists(ConnectorSession session,
 String schemaName)

Checks if a schema exists. The connector may have schemas that exist but are not enumerable via listSchemaNames(io.trino.spi.connector.ConnectorSession).

listSchemaNames

default List<String> listSchemaNames(ConnectorSession session)

Returns the schemas provided by this connector.

getTableHandle

@Nullable
@Deprecated
default ConnectorTableHandle getTableHandle(ConnectorSession session,
 SchemaTableName tableName)

Deprecated.

Implement getTableHandle(ConnectorSession, SchemaTableName, Optional, Optional).

Returns a table handle for the specified table name, or null if tableName relation does not exist or is not a table (e.g. is a view, or a materialized view).

Throws:

TrinoException - implementation can throw this exception when tableName refers to a table that cannot be queried.

See Also:

• getView(ConnectorSession, SchemaTableName)
• getMaterializedView(ConnectorSession, SchemaTableName)

getTableHandle

@Nullable
default ConnectorTableHandle getTableHandle(ConnectorSession session,
 SchemaTableName tableName,
 Optional<ConnectorTableVersion> startVersion,
 Optional<ConnectorTableVersion> endVersion)

Returns a table handle for the specified table name and version, or null if tableName relation does not exist or is not a table (e.g. is a view, or a materialized view).

Throws:

TrinoException - implementation can throw this exception when tableName refers to a table that cannot be queried.

See Also:

• getView(ConnectorSession, SchemaTableName)
• getMaterializedView(ConnectorSession, SchemaTableName)

getTableHandleForExecute

default Optional<ConnectorTableExecuteHandle> getTableHandleForExecute(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 String procedureName,
 Map<String,Object> executeProperties,
 RetryMode retryMode)

Create initial handle for execution of table procedure. The handle will be used through planning process. It will be converted to final handle used for execution via @{link beginTableExecute(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.ConnectorTableExecuteHandle, io.trino.spi.connector.ConnectorTableHandle)

If connector does not support execution with retries, the method should throw:

     new TrinoException(NOT_SUPPORTED, "This connector does not support query retries")
 

unless retryMode is set to NO_RETRIES.

getLayoutForTableExecute

default Optional<ConnectorTableLayout> getLayoutForTableExecute(ConnectorSession session,
 ConnectorTableExecuteHandle tableExecuteHandle)

beginTableExecute

default BeginTableExecuteResult<ConnectorTableExecuteHandle,ConnectorTableHandle> beginTableExecute(ConnectorSession session,
 ConnectorTableExecuteHandle tableExecuteHandle,
 ConnectorTableHandle updatedSourceTableHandle)

Begin execution of table procedure

finishTableExecute

default void finishTableExecute(ConnectorSession session,
 ConnectorTableExecuteHandle tableExecuteHandle,
 Collection<io.airlift.slice.Slice> fragments,
 List<Object> tableExecuteState)

Finish table execute

Parameters:

fragments - all fragments returned by ConnectorPageSink.finish()

executeTableExecute

default void executeTableExecute(ConnectorSession session,
 ConnectorTableExecuteHandle tableExecuteHandle)

Execute a coordinator-only table procedure.

getSystemTable

default Optional<SystemTable> getSystemTable(ConnectorSession session,
 SchemaTableName tableName)

Returns the system table for the specified table name, if one exists. The system tables handled via this method differ form those returned by Connector.getSystemTables(). The former mechanism allows dynamic resolution of system tables, while the latter is based on static list of system tables built during startup.

makeCompatiblePartitioning

default ConnectorTableHandle makeCompatiblePartitioning(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 ConnectorPartitioningHandle partitioningHandle)

Return a table handle whose partitioning is converted to the provided partitioning handle, but otherwise identical to the provided table handle. The provided table handle must be one that the connector can transparently convert to from the original partitioning handle associated with the provided table handle, as promised by getCommonPartitioningHandle(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.ConnectorPartitioningHandle, io.trino.spi.connector.ConnectorPartitioningHandle).

getCommonPartitioningHandle

default Optional<ConnectorPartitioningHandle> getCommonPartitioningHandle(ConnectorSession session,
 ConnectorPartitioningHandle left,
 ConnectorPartitioningHandle right)

Return a partitioning handle which the connector can transparently convert both left and right into.

getTableName

default SchemaTableName getTableName(ConnectorSession session,
 ConnectorTableHandle table)

Return schema table name for the specified table handle. This method is useful when requiring only SchemaTableName without other objects.

Throws:

RuntimeException - if table handle is no longer valid

getTableSchema

default ConnectorTableSchema getTableSchema(ConnectorSession session,
 ConnectorTableHandle table)

Return table schema definition for the specified table handle. This method is useful when getting full table metadata is expensive.

Throws:

RuntimeException - if table handle is no longer valid

getTableMetadata

default ConnectorTableMetadata getTableMetadata(ConnectorSession session,
 ConnectorTableHandle table)

Return the metadata for the specified table handle.

Throws:

RuntimeException - if table handle is no longer valid

getInfo

default Optional<Object> getInfo(ConnectorTableHandle table)

listTables

default List<SchemaTableName> listTables(ConnectorSession session,
 Optional<String> schemaName)

List table, view and materialized view names, possibly filtered by schema. An empty list is returned if none match. An empty list is returned also when schema name does not refer to an existing schema.

See Also:

• listViews(ConnectorSession, Optional)
• listMaterializedViews(ConnectorSession, Optional)

getColumnHandles

default Map<String,ColumnHandle> getColumnHandles(ConnectorSession session,
 ConnectorTableHandle tableHandle)

Gets all of the columns on the specified table, or an empty map if the columns cannot be enumerated.

Throws:

RuntimeException - if table handle is no longer valid

getColumnMetadata

default ColumnMetadata getColumnMetadata(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 ColumnHandle columnHandle)

Gets the metadata for the specified table column.

Throws:

RuntimeException - if table or column handles are no longer valid

listTableColumns

@Deprecated
default Map<SchemaTableName,List<ColumnMetadata>> listTableColumns(ConnectorSession session,
 SchemaTablePrefix prefix)

Deprecated.

use streamTableColumns(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.SchemaTablePrefix) which handles redirected tables

Gets the metadata for all columns that match the specified table prefix. Columns of views and materialized views are not included.

streamTableColumns

@Deprecated
default Iterator<TableColumnsMetadata> streamTableColumns(ConnectorSession session,
 SchemaTablePrefix prefix)

Deprecated.

Implement streamRelationColumns(io.trino.spi.connector.ConnectorSession, java.util.Optional<java.lang.String>, java.util.function.UnaryOperator<java.util.Set<io.trino.spi.connector.SchemaTableName>>).

Gets the metadata for all columns that match the specified table prefix. Redirected table names are included, but the column metadata for them is not. Views and materialized views are not included.

streamRelationColumns

@Experimental(eta="2024-01-01")
default Iterator<RelationColumnsMetadata> streamRelationColumns(ConnectorSession session,
 Optional<String> schemaName,
 UnaryOperator<Set<SchemaTableName>> relationFilter)

Gets columns for all relations (tables, views, materialized views), possibly filtered by schemaName. (e.g. for all relations that would be returned by listTables(ConnectorSession, Optional)). Redirected table names are included, but the comment for them is not.

streamRelationComments

@Experimental(eta="2024-01-01")
default Iterator<RelationCommentMetadata> streamRelationComments(ConnectorSession session,
 Optional<String> schemaName,
 UnaryOperator<Set<SchemaTableName>> relationFilter)

Gets comments for all relations (tables, views, materialized views), possibly filtered by schemaName. (e.g. for all relations that would be returned by listTables(ConnectorSession, Optional)). Redirected table names are included, but the comment for them is not.

getTableStatistics

default TableStatistics getTableStatistics(ConnectorSession session,
 ConnectorTableHandle tableHandle)

Get statistics for table.

createSchema

default void createSchema(ConnectorSession session,
 String schemaName,
 Map<String,Object> properties,
 TrinoPrincipal owner)

Creates a schema.

dropSchema

default void dropSchema(ConnectorSession session,
 String schemaName,
 boolean cascade)

Drops the specified schema.

Throws:

TrinoException - with SCHEMA_NOT_EMPTY if cascade is false and the schema is not empty

renameSchema

default void renameSchema(ConnectorSession session,
 String source,
 String target)

Renames the specified schema.

setSchemaAuthorization

default void setSchemaAuthorization(ConnectorSession session,
 String schemaName,
 TrinoPrincipal principal)

Sets the user/role on the specified schema.

createTable

@Deprecated
default void createTable(ConnectorSession session,
 ConnectorTableMetadata tableMetadata,
 boolean ignoreExisting)

Deprecated.

use createTable(ConnectorSession session, ConnectorTableMetadata tableMetadata, SaveMode saveMode)

Creates a table using the specified table metadata.

Throws:

TrinoException - with ALREADY_EXISTS if the table already exists and is not set

createTable

default void createTable(ConnectorSession session,
 ConnectorTableMetadata tableMetadata,
 SaveMode saveMode)

Creates a table using the specified table metadata. IGNORE means the table is created using CREATE ... IF NOT EXISTS syntax. REPLACE means the table is created using CREATE OR REPLACE syntax.

Throws:

TrinoException - with ALREADY_EXISTS if the table already exists and is FAIL.

dropTable

default void dropTable(ConnectorSession session,
 ConnectorTableHandle tableHandle)

Drops the specified table

Throws:

RuntimeException - if the table cannot be dropped or table handle is no longer valid

truncateTable

default void truncateTable(ConnectorSession session,
 ConnectorTableHandle tableHandle)

Truncates the specified table

Throws:

RuntimeException - if the table cannot be dropped or table handle is no longer valid

renameTable

default void renameTable(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 SchemaTableName newTableName)

Rename the specified table

setTableProperties

default void setTableProperties(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 Map<String,Optional<Object>> properties)

Set properties to the specified table

setTableComment

default void setTableComment(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 Optional<String> comment)

Comments to the specified table

setViewComment

default void setViewComment(ConnectorSession session,
 SchemaTableName viewName,
 Optional<String> comment)

Comments to the specified view

setViewColumnComment

default void setViewColumnComment(ConnectorSession session,
 SchemaTableName viewName,
 String columnName,
 Optional<String> comment)

Comments to the specified view column.

setMaterializedViewColumnComment

default void setMaterializedViewColumnComment(ConnectorSession session,
 SchemaTableName viewName,
 String columnName,
 Optional<String> comment)

Comments to the specified materialized view column.

setColumnComment

default void setColumnComment(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 ColumnHandle column,
 Optional<String> comment)

Comments to the specified column

addColumn

default void addColumn(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 ColumnMetadata column)

Add the specified column

addField

@Experimental(eta="2023-06-01")
default void addField(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 List<String> parentPath,
 String fieldName,
 Type type,
 boolean ignoreExisting)

Add the specified field, potentially nested, to a row.

Parameters:

parentPath - path to a field within the column, without leaf field name.

setColumnType

@Experimental(eta="2023-04-01")
default void setColumnType(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 ColumnHandle column,
 Type type)

Set the specified column type

setFieldType

@Experimental(eta="2023-09-01")
default void setFieldType(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 List<String> fieldPath,
 Type type)

Set the specified field type

Parameters:

fieldPath - path starting with column name. The path is always lower-cased. It cannot be an empty or a single element.

setTableAuthorization

default void setTableAuthorization(ConnectorSession session,
 SchemaTableName tableName,
 TrinoPrincipal principal)

Sets the user/role on the specified table.

renameColumn

default void renameColumn(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 ColumnHandle source,
 String target)

Rename the specified column

renameField

@Experimental(eta="2023-09-01")
default void renameField(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 List<String> fieldPath,
 String target)

Rename the specified field, potentially nested, to a row.

Parameters:

fieldPath - path starting with column name.

target - the new field name. The field position and nested level shouldn't be changed.

dropColumn

default void dropColumn(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 ColumnHandle column)

Drop the specified column

dropField

@Experimental(eta="2023-05-01")
default void dropField(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 ColumnHandle column,
 List<String> fieldPath)

Drop the specified field, potentially nested, from a row.

Parameters:

fieldPath - path to a field within the column, without leading column name.

getNewTableLayout

default Optional<ConnectorTableLayout> getNewTableLayout(ConnectorSession session,
 ConnectorTableMetadata tableMetadata)

Get the physical layout for a new table.

getSupportedType

@Experimental(eta="2024-01-31")
default Optional<Type> getSupportedType(ConnectorSession session,
 Map<String,Object> tableProperties,
 Type type)

Return the effective Type that is supported by the connector for the given type. If Optional.empty() is returned, the type will be used as is during table creation which may or may not be supported by the connector. The effective type shall be a type that is cast-compatible with the input type.

getInsertLayout

default Optional<ConnectorTableLayout> getInsertLayout(ConnectorSession session,
 ConnectorTableHandle tableHandle)

Get the physical layout for inserting into an existing table.

getStatisticsCollectionMetadataForWrite

default TableStatisticsMetadata getStatisticsCollectionMetadataForWrite(ConnectorSession session,
 ConnectorTableMetadata tableMetadata)

Describes statistics that must be collected during a write.

getStatisticsCollectionMetadata

default ConnectorAnalyzeMetadata getStatisticsCollectionMetadata(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 Map<String,Object> analyzeProperties)

Describe statistics that must be collected during a statistics collection

beginStatisticsCollection

default ConnectorTableHandle beginStatisticsCollection(ConnectorSession session,
 ConnectorTableHandle tableHandle)

Begin statistics collection

finishStatisticsCollection

default void finishStatisticsCollection(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 Collection<ComputedStatistics> computedStatistics)

Finish statistics collection

beginCreateTable

@Deprecated
default ConnectorOutputTableHandle beginCreateTable(ConnectorSession session,
 ConnectorTableMetadata tableMetadata,
 Optional<ConnectorTableLayout> layout,
 RetryMode retryMode)

Deprecated.

use beginCreateTable(ConnectorSession session, ConnectorTableMetadata tableMetadata, Optional layout, RetryMode retryMode, boolean replace)

Begin the atomic creation of a table with data.

If connector does not support execution with retries, the method should throw:

     new TrinoException(NOT_SUPPORTED, "This connector does not support query retries")
 

unless retryMode is set to NO_RETRIES.

beginCreateTable

default ConnectorOutputTableHandle beginCreateTable(ConnectorSession session,
 ConnectorTableMetadata tableMetadata,
 Optional<ConnectorTableLayout> layout,
 RetryMode retryMode,
 boolean replace)

Begin the atomic creation of a table with data.

If connector does not support execution with retries, the method should throw:

     new TrinoException(NOT_SUPPORTED, "This connector does not support query retries")
 

unless retryMode is set to NO_RETRIES.

finishCreateTable

default Optional<ConnectorOutputMetadata> finishCreateTable(ConnectorSession session,
 ConnectorOutputTableHandle tableHandle,
 Collection<io.airlift.slice.Slice> fragments,
 Collection<ComputedStatistics> computedStatistics)

Finish a table creation with data after the data is written.

beginQuery

default void beginQuery(ConnectorSession session)

Start a query. This notification is triggered before any other metadata access.

cleanupQuery

default void cleanupQuery(ConnectorSession session)

Cleanup after a query. This is the very last notification after the query finishes, whether it succeeds or fails. An exception thrown in this method will not affect the result of the query.

beginInsert

default ConnectorInsertTableHandle beginInsert(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 List<ColumnHandle> columns,
 RetryMode retryMode)

Begin insert query.

If connector does not support execution with retries, the method should throw:

     new TrinoException(NOT_SUPPORTED, "This connector does not support query retries")
 

unless retryMode is set to NO_RETRIES.

supportsMissingColumnsOnInsert

default boolean supportsMissingColumnsOnInsert()

Returns:

whether connector handles missing columns during insert

finishInsert

default Optional<ConnectorOutputMetadata> finishInsert(ConnectorSession session,
 ConnectorInsertTableHandle insertHandle,
 Collection<io.airlift.slice.Slice> fragments,
 Collection<ComputedStatistics> computedStatistics)

Finish insert query

delegateMaterializedViewRefreshToConnector

default boolean delegateMaterializedViewRefreshToConnector(ConnectorSession session,
 SchemaTableName viewName)

Returns true if materialized view refresh should be delegated to connector using refreshMaterializedView(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.SchemaTableName)

refreshMaterializedView

default CompletableFuture<?> refreshMaterializedView(ConnectorSession session,
 SchemaTableName viewName)

Refresh materialized view

beginRefreshMaterializedView

default ConnectorInsertTableHandle beginRefreshMaterializedView(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 List<ConnectorTableHandle> sourceTableHandles,
 RetryMode retryMode)

Begin materialized view query.

If connector does not support execution with retries, the method should throw:

     new TrinoException(NOT_SUPPORTED, "This connector does not support query retries")
 

unless retryMode is set to NO_RETRIES.

finishRefreshMaterializedView

default Optional<ConnectorOutputMetadata> finishRefreshMaterializedView(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 ConnectorInsertTableHandle insertHandle,
 Collection<io.airlift.slice.Slice> fragments,
 Collection<ComputedStatistics> computedStatistics,
 List<ConnectorTableHandle> sourceTableHandles)

Finish materialized view query

getRowChangeParadigm

default RowChangeParadigm getRowChangeParadigm(ConnectorSession session,
 ConnectorTableHandle tableHandle)

Return the row change paradigm supported by the connector on the table.

getMergeRowIdColumnHandle

default ColumnHandle getMergeRowIdColumnHandle(ConnectorSession session,
 ConnectorTableHandle tableHandle)

Get the column handle that will generate row IDs for the merge operation. These IDs will be passed to the ConnectorMergeSink.storeMergedRows(io.trino.spi.Page) method of the ConnectorMergeSink that created them.

getUpdateLayout

default Optional<ConnectorPartitioningHandle> getUpdateLayout(ConnectorSession session,
 ConnectorTableHandle tableHandle)

Get the physical layout for updated or deleted rows of a MERGE operation. Inserted rows are handled by getInsertLayout(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.ConnectorTableHandle). This layout always uses the merge row ID column.

beginMerge

default ConnectorMergeTableHandle beginMerge(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 RetryMode retryMode)

Do whatever is necessary to start an MERGE query, returning the ConnectorMergeTableHandle instance that will be passed to the PageSink, and to the finishMerge(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.ConnectorMergeTableHandle, java.util.Collection<io.airlift.slice.Slice>, java.util.Collection<io.trino.spi.statistics.ComputedStatistics>) method.

finishMerge

default void finishMerge(ConnectorSession session,
 ConnectorMergeTableHandle mergeTableHandle,
 Collection<io.airlift.slice.Slice> fragments,
 Collection<ComputedStatistics> computedStatistics)

Finish a merge query

Parameters:

session - The session

mergeTableHandle - A ConnectorMergeTableHandle for the table that is the target of the merge

fragments - All fragments returned by the merge plan

computedStatistics - Statistics for the table, meaningful only to the connector that produced them.

createView

default void createView(ConnectorSession session,
 SchemaTableName viewName,
 ConnectorViewDefinition definition,
 boolean replace)

Create the specified view. The view definition is intended to be serialized by the connector for permanent storage.

renameView

default void renameView(ConnectorSession session,
 SchemaTableName source,
 SchemaTableName target)

Rename the specified view

setViewAuthorization

default void setViewAuthorization(ConnectorSession session,
 SchemaTableName viewName,
 TrinoPrincipal principal)

Sets the user/role on the specified view.

dropView

default void dropView(ConnectorSession session,
 SchemaTableName viewName)

Drop the specified view.

listViews

default List<SchemaTableName> listViews(ConnectorSession session,
 Optional<String> schemaName)

List view names (but not materialized views), possibly filtered by schema. An empty list is returned if none match. An empty list is returned also when schema name does not refer to an existing schema.

See Also:

• listMaterializedViews(ConnectorSession, Optional)

getViews

default Map<SchemaTableName,ConnectorViewDefinition> getViews(ConnectorSession session,
 Optional<String> schemaName)

Gets the definitions of views (but not materialized views), possibly filtered by schema. This optional method may be implemented by connectors that can support fetching view data in bulk. It is used to implement information_schema.views.

getView

default Optional<ConnectorViewDefinition> getView(ConnectorSession session,
 SchemaTableName viewName)

Gets the view data for the specified view name. Returns Optional.empty() if viewName relation does not or is not a view (e.g. is a table, or a materialized view).

See Also:

• getTableHandle(ConnectorSession, SchemaTableName)
• getMaterializedView(ConnectorSession, SchemaTableName)

getSchemaProperties

default Map<String,Object> getSchemaProperties(ConnectorSession session,
 String schemaName)

Gets the schema properties for the specified schema.

getSchemaOwner

default Optional<TrinoPrincipal> getSchemaOwner(ConnectorSession session,
 String schemaName)

Get the schema properties for the specified schema.

applyUpdate

default Optional<ConnectorTableHandle> applyUpdate(ConnectorSession session,
 ConnectorTableHandle handle,
 Map<ColumnHandle,Constant> assignments)

Attempt to push down an update operation into the connector. If a connector can execute an update for the table handle on its own, it should return a table handle, which will be passed back to executeUpdate(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.ConnectorTableHandle) during query executing to actually execute the update.

executeUpdate

default OptionalLong executeUpdate(ConnectorSession session,
 ConnectorTableHandle handle)

Execute the update operation on the handle returned from applyUpdate(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.ConnectorTableHandle, java.util.Map<io.trino.spi.connector.ColumnHandle, io.trino.spi.expression.Constant>).

applyDelete

default Optional<ConnectorTableHandle> applyDelete(ConnectorSession session,
 ConnectorTableHandle handle)

Attempt to push down a delete operation into the connector. If a connector can execute a delete for the table handle on its own, it should return a table handle, which will be passed back to executeDelete(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.ConnectorTableHandle) during query executing to actually execute the delete.

executeDelete

default OptionalLong executeDelete(ConnectorSession session,
 ConnectorTableHandle handle)

Execute the delete operation on the handle returned from applyDelete(io.trino.spi.connector.ConnectorSession, io.trino.spi.connector.ConnectorTableHandle).

resolveIndex

default Optional<ConnectorResolvedIndex> resolveIndex(ConnectorSession session,
 ConnectorTableHandle tableHandle,
 Set<ColumnHandle> indexableColumns,
 Set<ColumnHandle> outputColumns,
 TupleDomain<ColumnHandle> tupleDomain)

Try to locate a table index that can lookup results by indexableColumns and provide the requested outputColumns.

listFunctions

default Collection<FunctionMetadata> listFunctions(ConnectorSession session,
 String schemaName)

List available functions.

getFunctions

default Collection<FunctionMetadata> getFunctions(ConnectorSession session,
 SchemaFunctionName name)

Get all functions with specified name.

getFunctionMetadata

default FunctionMetadata getFunctionMetadata(ConnectorSession session,
 FunctionId functionId)

Return the function with the specified id.

getAggregationFunctionMetadata

default AggregationFunctionMetadata getAggregationFunctionMetadata(ConnectorSession session,
 FunctionId functionId)

Returns the aggregation metadata for the aggregation function with the specified id.

getFunctionDependencies

default FunctionDependencyDeclaration getFunctionDependencies(ConnectorSession session,
 FunctionId functionId,
 BoundSignature boundSignature)

Returns the dependencies of the function with the specified id.

listLanguageFunctions

default Collection<LanguageFunction> listLanguageFunctions(ConnectorSession session,
 String schemaName)

List available language functions.

getLanguageFunctions

default Collection<LanguageFunction> getLanguageFunctions(ConnectorSession session,
 SchemaFunctionName name)

Get all language functions with the specified name.

languageFunctionExists

default boolean languageFunctionExists(ConnectorSession session,
 SchemaFunctionName name,
 String signatureToken)

Check if a language function exists.

createLanguageFunction

default void createLanguageFunction(ConnectorSession session,
 SchemaFunctionName name,
 LanguageFunction function,
 boolean replace)

Creates a language function with the specified name and signature token. The signature token is an opaque string that uniquely identifies the function signature. Multiple functions with the same name but with different signatures may exist. The signature token is used to identify the function when dropping it.

Parameters:

replace - if true, replace existing function with the same name and signature token

dropLanguageFunction

default void dropLanguageFunction(ConnectorSession session,
 SchemaFunctionName name,
 String signatureToken)

Drops a language function with the specified name and signature token.

roleExists

default boolean roleExists(ConnectorSession session,
 String role)

Does the specified role exist.

createRole

default void createRole(ConnectorSession session,
 String role,
 Optional<TrinoPrincipal> grantor)

Creates the specified role.

Parameters:

grantor - represents the principal specified by WITH ADMIN statement

dropRole

default void dropRole(ConnectorSession session,
 String role)

Drops the specified role.

listRoles

default Set<String> listRoles(ConnectorSession session)

List available roles.

listRoleGrants

default Set<RoleGrant> listRoleGrants(ConnectorSession session,
 TrinoPrincipal principal)

List role grants for a given principal, not recursively.

grantRoles

default void grantRoles(ConnectorSession connectorSession,
 Set<String> roles,
 Set<TrinoPrincipal> grantees,
 boolean adminOption,
 Optional<TrinoPrincipal> grantor)

Grants the specified roles to the specified grantees

Parameters:

grantor - represents the principal specified by GRANTED BY statement

revokeRoles

default void revokeRoles(ConnectorSession connectorSession,
 Set<String> roles,
 Set<TrinoPrincipal> grantees,
 boolean adminOption,
 Optional<TrinoPrincipal> grantor)

Revokes the specified roles from the specified grantees

Parameters:

grantor - represents the principal specified by GRANTED BY statement

listApplicableRoles

default Set<RoleGrant> listApplicableRoles(ConnectorSession session,
 TrinoPrincipal principal)

List applicable roles, including the transitive grants, for the specified principal

listEnabledRoles

default Set<String> listEnabledRoles(ConnectorSession session)

List applicable roles, including the transitive grants, in given session

grantSchemaPrivileges

default void grantSchemaPrivileges(ConnectorSession session,
 String schemaName,
 Set<Privilege> privileges,
 TrinoPrincipal grantee,
 boolean grantOption)

Grants the specified privilege to the specified user on the specified schema

denySchemaPrivileges

default void denySchemaPrivileges(ConnectorSession session,
 String schemaName,
 Set<Privilege> privileges,
 TrinoPrincipal grantee)

Denys the specified privilege to the specified user on the specified schema

revokeSchemaPrivileges

default void revokeSchemaPrivileges(ConnectorSession session,
 String schemaName,
 Set<Privilege> privileges,
 TrinoPrincipal grantee,
 boolean grantOption)

Revokes the specified privilege on the specified schema from the specified user

grantTablePrivileges

default void grantTablePrivileges(ConnectorSession session,
 SchemaTableName tableName,
 Set<Privilege> privileges,
 TrinoPrincipal grantee,
 boolean grantOption)

Grants the specified privilege to the specified user on the specified table

denyTablePrivileges

default void denyTablePrivileges(ConnectorSession session,
 SchemaTableName tableName,
 Set<Privilege> privileges,
 TrinoPrincipal grantee)

Denys the specified privilege to the specified user on the specified table

revokeTablePrivileges

default void revokeTablePrivileges(ConnectorSession session,
 SchemaTableName tableName,
 Set<Privilege> privileges,
 TrinoPrincipal grantee,
 boolean grantOption)

Revokes the specified privilege on the specified table from the specified user

listTablePrivileges

default List<GrantInfo> listTablePrivileges(ConnectorSession session,
 SchemaTablePrefix prefix)

List the table privileges granted to the specified grantee for the tables that have the specified prefix considering the selected session role

getTableProperties

default ConnectorTableProperties getTableProperties(ConnectorSession session,
 ConnectorTableHandle table)

applyLimit

default Optional<LimitApplicationResult<ConnectorTableHandle>> applyLimit(ConnectorSession session,
 ConnectorTableHandle handle,
 long limit)

Attempt to push down the provided limit into the table.

Connectors can indicate whether they don't support limit pushdown or that the action had no effect by returning Optional.empty(). Connectors should expect this method to be called multiple times during the optimization of a given query.

Note: it's critical for connectors to return Optional.empty() if calling this method has no effect for that invocation, even if the connector generally supports limit pushdown. Doing otherwise can cause the optimizer to loop indefinitely.

If the connector could benefit from the information but can't guarantee that it will be able to produce fewer rows than the provided limit, it should return a non-empty result containing a new handle for the derived table and the "limit guaranteed" flag set to false.

If the connector can guarantee it will produce fewer rows than the provided limit, it should return a non-empty result with the "limit guaranteed" flag set to true.

applyFilter

default Optional<ConstraintApplicationResult<ConnectorTableHandle>> applyFilter(ConnectorSession session,
 ConnectorTableHandle handle,
 Constraint constraint)

Attempt to push down the provided constraint into the table.

Connectors can indicate whether they don't support predicate pushdown or that the action had no effect by returning Optional.empty(). Connectors should expect this method to be called multiple times during the optimization of a given query.

Note: it's critical for connectors to return Optional.empty() if calling this method has no effect for that invocation, even if the connector generally supports pushdown. Doing otherwise can cause the optimizer to loop indefinitely.

Note: Implementation must not maintain reference to constraint's Constraint.predicate() after the call returns.

Parameters:

constraint - constraint to be applied to the table. Constraint.getSummary() is guaranteed not to be none.

applyProjection

default Optional<ProjectionApplicationResult<ConnectorTableHandle>> applyProjection(ConnectorSession session,
 ConnectorTableHandle handle,
 List<ConnectorExpression> projections,
 Map<String,ColumnHandle> assignments)

Attempt to push down the provided projections into the table.

Connectors can indicate whether they don't support projection pushdown or that the action had no effect by returning Optional.empty(). Connectors should expect this method to be called multiple times during the optimization of a given query.

Note: it's critical for connectors to return Optional.empty() if calling this method has no effect for that invocation, even if the connector generally supports pushdown. Doing otherwise can cause the optimizer to loop indefinitely.

If the method returns a result, the list of projections in the result *replaces* the existing ones, and the list of assignments is the new set of columns exposed by the derived table.

As an example, given the following plan:

 - project
     x = f1(a, b)
     y = f2(a, b)
     z = f3(a, b)
   - scan (TH0)
       a = CH0
       b = CH1
       c = CH2
 

The optimizer would call this method with the following arguments:

 handle = TH0
 projections = [
     f1(a, b)
     f2(a, b)
     f3(a, b)
 ]
 assignments = [
     a = CH0
     b = CH1
     c = CH2
 ]
 

Assuming the connector knows how to handle f1(...) and f2(...), it would return:

 handle = TH1
 projections = [
     v2
     v3
     f3(v0, v1)
 ]
 assignments = [
     v0 = CH0
     v1 = CH1
     v2 = CH3  (synthetic column for f1(CH0, CH1))
     v3 = CH4  (synthetic column for f2(CH0, CH1))
 ]
 

applySample

default Optional<SampleApplicationResult<ConnectorTableHandle>> applySample(ConnectorSession session,
 ConnectorTableHandle handle,
 SampleType sampleType,
 double sampleRatio)

Attempt to push down the sampling into the table.

Connectors can indicate whether they don't support sample pushdown or that the action had no effect by returning Optional.empty(). Connectors should expect this method to be called multiple times during the optimization of a given query.

Note: it's critical for connectors to return Optional.empty() if calling this method has no effect for that invocation, even if the connector generally supports sample pushdown. Doing otherwise can cause the optimizer to loop indefinitely.

applyAggregation

default Optional<AggregationApplicationResult<ConnectorTableHandle>> applyAggregation(ConnectorSession session,
 ConnectorTableHandle handle,
 List<AggregateFunction> aggregates,
 Map<String,ColumnHandle> assignments,
 List<List<ColumnHandle>> groupingSets)

Attempt to push down the aggregates into the table.

Connectors can indicate whether they don't support aggregate pushdown or that the action had no effect by returning Optional.empty(). Connectors should expect this method may be called multiple times.

Note: it's critical for connectors to return Optional.empty() if calling this method has no effect for that invocation, even if the connector generally supports pushdown. Doing otherwise can cause the optimizer to loop indefinitely.

If the method returns a result, the list of assignments in the result will be merged with existing assignments. The projections returned by the method must have the same order as the given input list of aggregates.

As an example, given the following plan:

  - aggregation  (GROUP BY c)
          variable0 = agg_fn1(a)
          variable1 = agg_fn2(b, 2)
          variable2 = c
          - scan (TH0)
              a = CH0
              b = CH1
              c = CH2
 

The optimizer would call this method with the following arguments:

      handle = TH0
      aggregates = [
              { functionName=agg_fn1, outputType = «some Trino type» inputs = [Variable a]} ,
              { functionName=agg_fn2, outputType = «some Trino type» inputs = [Variable b, Constant 2]}
      ]
      groupingSets=[[ColumnHandle CH2]]
      assignments = {a = CH0, b = CH1, c = CH2}
 

Assuming the connector knows how to handle agg_fn1(...) and agg_fn2(...), it would return:

 AggregationApplicationResult {
      handle = TH1
      projections = [Variable synthetic_name0, Variable synthetic_name1] -- The order in the list must be same as input list of aggregates
      assignments = {
          synthetic_name0 = CH3 (synthetic column for agg_fn1(a))
          synthetic_name1 = CH4 (synthetic column for agg_fn2(b,2))
      }
 }
 

if the connector only knows how to handle agg_fn1(...), but not agg_fn2, it should return Optional.empty().

Another example is where the connector wants to handle the aggregate function by pointing to a pre-materialized table. In this case the input can stay same as in the above example and the connector can return

 AggregationApplicationResult {
      handle = TH1 (could capture information about which pre-materialized table to use)
      projections = [Variable synthetic_name0, Variable synthetic_name1] -- The order in the list must be same as input list of aggregates
      assignments = {
          synthetic_name0 = CH3 (reference to the column in pre-materialized table that has agg_fn1(a) calculated)
          synthetic_name1 = CH4 (reference to the column in pre-materialized table that has agg_fn2(b,2) calculated)
          synthetic_name2 = CH5 (reference to the column in pre-materialized table that has the group by column c)
      }
      groupingColumnMapping = {
          CH2 -> CH5 (CH2 in the original assignment should now be replaced by CH5 in the new assignment)
      }
 }
 

applyJoin

default Optional<JoinApplicationResult<ConnectorTableHandle>> applyJoin(ConnectorSession session,
 JoinType joinType,
 ConnectorTableHandle left,
 ConnectorTableHandle right,
 ConnectorExpression joinCondition,
 Map<String,ColumnHandle> leftAssignments,
 Map<String,ColumnHandle> rightAssignments,
 JoinStatistics statistics)

Attempt to push down the join operation.

Connectors can indicate whether they don't support join pushdown or that the action had no effect by returning Optional.empty(). Connectors should expect this method may be called multiple times.

Warning: this is an experimental API and it will change in the future.

Join condition conjuncts are passed via joinConditions list. For current implementation connector may assume that leftExpression and rightExpression in each of the conjucts are instances of Variable. This may be relaxed in the future.

The leftAssignments and rightAssignments lists provide mappings from variable names, used in joinConditions to input tables column handles. It is guaranteed that all the required mappings will be provided but not necessarily *all* the column handles of tables which are join inputs.

Table statistics for left, right table as well as estimated statistics for join are provided via statistics parameter. Those can be used by connector to assess if performing join pushdown is expected to improve query performance.

If the method returns a result the returned table handle will be used in place of join and input table scans. Returned result must provide mapping from old column handles to new ones via leftColumnHandles and rightColumnHandles fields of the result. It is required that mapping is provided for *all* column handles exposed previously by both left and right join sources.

applyJoin

@Deprecated
default Optional<JoinApplicationResult<ConnectorTableHandle>> applyJoin(ConnectorSession session,
 JoinType joinType,
 ConnectorTableHandle left,
 ConnectorTableHandle right,
 List<JoinCondition> joinConditions,
 Map<String,ColumnHandle> leftAssignments,
 Map<String,ColumnHandle> rightAssignments,
 JoinStatistics statistics)

Deprecated.

applyTopN

default Optional<TopNApplicationResult<ConnectorTableHandle>> applyTopN(ConnectorSession session,
 ConnectorTableHandle handle,
 long topNCount,
 List<SortItem> sortItems,
 Map<String,ColumnHandle> assignments)

Attempt to push down the TopN into the table scan.

Connectors can indicate whether they don't support topN pushdown or that the action had no effect by returning Optional.empty(). Connectors should expect this method may be called multiple times.

Note: it's critical for connectors to return Optional.empty() if calling this method has no effect for that invocation, even if the connector generally supports topN pushdown. Doing otherwise can cause the optimizer to loop indefinitely.

If the connector can handle TopN Pushdown and guarantee it will produce no more rows than requested then it should return a non-empty result with "topN guaranteed" flag set to true.

applyTableFunction

default Optional<TableFunctionApplicationResult<ConnectorTableHandle>> applyTableFunction(ConnectorSession session,
 ConnectorTableFunctionHandle handle)

Attempt to push down the table function invocation into the connector.

Connectors can indicate whether they don't support table function invocation pushdown or that the action had no effect by returning Optional.empty(). Connectors should expect this method may be called multiple times.

If the method returns a result, the returned table handle will be used in place of the table function invocation.

validateScan

default void validateScan(ConnectorSession session,
 ConnectorTableHandle handle)

Allows the connector to reject the table scan produced by the planner.

Connectors can choose to reject a query based on the table scan potentially being too expensive, for example if no filtering is done on a partition column.

createMaterializedView

default void createMaterializedView(ConnectorSession session,
 SchemaTableName viewName,
 ConnectorMaterializedViewDefinition definition,
 boolean replace,
 boolean ignoreExisting)

Create the specified materialized view. The view definition is intended to be serialized by the connector for permanent storage.

Throws:

TrinoException - with ALREADY_EXISTS if the object already exists and is not set

dropMaterializedView

default void dropMaterializedView(ConnectorSession session,
 SchemaTableName viewName)

Drop the specified materialized view.

listMaterializedViews

default List<SchemaTableName> listMaterializedViews(ConnectorSession session,
 Optional<String> schemaName)

List materialized view names, possibly filtered by schema. An empty list is returned if none match. An empty list is returned also when schema name does not refer to an existing schema.

getMaterializedViews

default Map<SchemaTableName,ConnectorMaterializedViewDefinition> getMaterializedViews(ConnectorSession session,
 Optional<String> schemaName)

Gets the definitions of materialized views, possibly filtered by schema. This optional method may be implemented by connectors that can support fetching view data in bulk. It is used to populate information_schema.columns.

getMaterializedView

default Optional<ConnectorMaterializedViewDefinition> getMaterializedView(ConnectorSession session,
 SchemaTableName viewName)

Gets the materialized view data for the specified materialized view name. Returns Optional.empty() if viewName relation does not or is not a materialized view (e.g. is a table, or a view).

See Also:

• getTableHandle(ConnectorSession, SchemaTableName)
• getView(ConnectorSession, SchemaTableName)

getMaterializedViewFreshness

default MaterializedViewFreshness getMaterializedViewFreshness(ConnectorSession session,
 SchemaTableName name)

The method is used by the engine to determine if a materialized view is current with respect to the tables it depends on.

Throws:

MaterializedViewNotFoundException - when materialized view is not found

renameMaterializedView

default void renameMaterializedView(ConnectorSession session,
 SchemaTableName source,
 SchemaTableName target)

Rename the specified materialized view

setMaterializedViewProperties

default void setMaterializedViewProperties(ConnectorSession session,
 SchemaTableName viewName,
 Map<String,Optional<Object>> properties)

Sets the properties of the specified materialized view

applyTableScanRedirect

default Optional<TableScanRedirectApplicationResult> applyTableScanRedirect(ConnectorSession session,
 ConnectorTableHandle tableHandle)

redirectTable

default Optional<CatalogSchemaTableName> redirectTable(ConnectorSession session,
 SchemaTableName tableName)

Redirects table to other table which may or may not be in the same catalog. Currently the engine tries to do redirection only for table reads and metadata listing.

Also consider implementing streamTableColumns to support redirection for listing.

getMaxWriterTasks

default OptionalInt getMaxWriterTasks(ConnectorSession session)

getNewTableWriterScalingOptions

default WriterScalingOptions getNewTableWriterScalingOptions(ConnectorSession session,
 SchemaTableName tableName,
 Map<String,Object> tableProperties)

getInsertWriterScalingOptions

default WriterScalingOptions getInsertWriterScalingOptions(ConnectorSession session,
 ConnectorTableHandle tableHandle)