Deprecated API

Contents

• Interfaces
• Classes
• Exceptions
• Fields
• Methods
• Constructors

Interfaces

Interface	Description
org.apache.flink.table.catalog.CatalogLock	This interface will be removed soon. Please see FLIP-346 for more details.
org.apache.flink.table.catalog.CatalogLock.Factory	This interface will be removed soon. Please see FLIP-346 for more details.
org.apache.flink.table.connector.RequireCatalogLock	This interface will be removed soon. Please see FLIP-346 for more details.
org.apache.flink.table.descriptors.DescriptorValidator	See Descriptor for details.
org.apache.flink.table.factories.ManagedTableFactory	This interface will be removed soon. Please see FLIP-346 for more details.
org.apache.flink.table.legacy.api.constraints.Constraint	See ResolvedSchema and Constraint.
org.apache.flink.table.legacy.connector.source.AsyncTableFunctionProvider	Please use AsyncLookupFunctionProvider to implement asynchronous lookup table.
org.apache.flink.table.legacy.connector.source.TableFunctionProvider	Please use LookupFunctionProvider to implement synchronous lookup table.
org.apache.flink.table.legacy.descriptors.Descriptor	Descriptor was primarily used for the legacy connector stack and have been deprecated. Use TableDescriptor for creating sources and sinks from the Table API.
org.apache.flink.table.legacy.factories.TableFactory	This interface has been replaced by Factory.
org.apache.flink.table.legacy.factories.TableSinkFactory	This interface has been replaced by DynamicTableSinkFactory. The new interface consumes internal data structures. See FLIP-95 for more information.
org.apache.flink.table.legacy.factories.TableSourceFactory	This interface has been replaced by DynamicTableSourceFactory. The new interface produces internal data structures. See FLIP-95 for more information.
org.apache.flink.table.legacy.sinks.OverwritableTableSink	This interface will not be supported in the new sink design around DynamicTableSink. Use SupportsOverwrite instead. See FLIP-95 for more information.
org.apache.flink.table.legacy.sinks.PartitionableTableSink	This interface will not be supported in the new sink design around DynamicTableSink. Use SupportsPartitioning instead. See FLIP-95 for more information.
org.apache.flink.table.legacy.sinks.TableSink	This interface has been replaced by DynamicTableSink. The new interface consumes internal data structures. See FLIP-95 for more information.
org.apache.flink.table.legacy.sources.DefinedFieldMapping	This interface will not be supported in the new source design around DynamicTableSource. See FLIP-95 for more information.
org.apache.flink.table.legacy.sources.DefinedProctimeAttribute	This interface will not be supported in the new source design around DynamicTableSource. Use the concept of computed columns instead. See FLIP-95 for more information.
org.apache.flink.table.legacy.sources.DefinedRowtimeAttributes	This interface will not be supported in the new source design around DynamicTableSource. Use the concept of computed columns instead. See FLIP-95 for more information.
org.apache.flink.table.legacy.sources.FieldComputer	This interface will not be supported in the new source design around DynamicTableSource. Use the concept of computed columns instead. See FLIP-95 for more information.
org.apache.flink.table.legacy.sources.FilterableTableSource	This interface will not be supported in the new source design around DynamicTableSource. Use SupportsFilterPushDown instead. See FLIP-95 for more information.
org.apache.flink.table.legacy.sources.LimitableTableSource	This interface will not be supported in the new source design around DynamicTableSource. Use SupportsLimitPushDown instead. See FLIP-95 for more information.
org.apache.flink.table.legacy.sources.LookupableTableSource	This interface will not be supported in the new source design around DynamicTableSource. Use LookupTableSource instead. See FLIP-95 for more information.
org.apache.flink.table.legacy.sources.NestedFieldsProjectableTableSource	This interface will not be supported in the new source design around DynamicTableSource. Use SupportsProjectionPushDown instead. See FLIP-95 for more information.
org.apache.flink.table.legacy.sources.PartitionableTableSource	This interface will not be supported in the new source design around DynamicTableSource. Use SupportsPartitionPushDown instead. See FLIP-95 for more information.
org.apache.flink.table.legacy.sources.ProjectableTableSource	This interface will not be supported in the new source design around DynamicTableSource. Use SupportsProjectionPushDown instead. See FLIP-95 for more information.
org.apache.flink.table.legacy.sources.TableSource	This interface has been replaced by DynamicTableSource. The new interface produces internal data structures. See FLIP-95 for more information.

Classes

Class	Description
org.apache.flink.table.dataview.ListViewSerializer
org.apache.flink.table.dataview.ListViewSerializerSnapshot
org.apache.flink.table.dataview.ListViewTypeInfo
org.apache.flink.table.dataview.ListViewTypeInfoFactory
org.apache.flink.table.dataview.MapViewSerializer
org.apache.flink.table.dataview.MapViewSerializerSnapshot
org.apache.flink.table.dataview.MapViewTypeInfo
org.apache.flink.table.dataview.MapViewTypeInfoFactory
org.apache.flink.table.dataview.NullAwareMapSerializer
org.apache.flink.table.dataview.NullAwareMapSerializerSnapshot
org.apache.flink.table.descriptors.ConnectorDescriptorValidator
org.apache.flink.table.descriptors.DescriptorProperties	This utility will be dropped soon. DynamicTableFactory is based on ConfigOption and catalogs use CatalogPropertiesUtil.
org.apache.flink.table.descriptors.FileSystemValidator	The legacy CSV connector has been replaced by FileSource / FileSink. It is kept only to support tests for the legacy connector stack.
org.apache.flink.table.factories.TableFactoryService
org.apache.flink.table.factories.TableSinkFactoryContextImpl
org.apache.flink.table.factories.TableSourceFactoryContextImpl
org.apache.flink.table.functions.AggregateFunctionDefinition	Non-legacy functions can simply omit this wrapper for declarations.
org.apache.flink.table.functions.LegacyUserDefinedFunctionInference
org.apache.flink.table.functions.ScalarFunctionDefinition	Non-legacy functions can simply omit this wrapper for declarations.
org.apache.flink.table.functions.TableAggregateFunctionDefinition	Non-legacy functions can simply omit this wrapper for declarations.
org.apache.flink.table.functions.TableFunctionDefinition	Non-legacy functions can simply omit this wrapper for declarations.
org.apache.flink.table.legacy.api.constraints.UniqueConstraint	See ResolvedSchema and UniqueConstraint.
org.apache.flink.table.legacy.api.TableColumn	See ResolvedSchema and Column.
org.apache.flink.table.legacy.api.TableSchema	This class has been deprecated as part of FLIP-164. It has been replaced by two more dedicated classes Schema and ResolvedSchema. Use Schema for declaration in APIs. ResolvedSchema is offered by the framework after resolution and validation.
org.apache.flink.table.legacy.api.Types	This class will be removed in future versions as it uses the old type system. It is recommended to use DataTypes instead which uses the new type system based on instances of DataType. Please make sure to use either the old or the new type system consistently to avoid unintended behavior. See the website documentation for more information.
org.apache.flink.table.legacy.api.WatermarkSpec	See ResolvedSchema and WatermarkSpec.
org.apache.flink.table.legacy.descriptors.Rowtime	This class was used for legacy connectors using Descriptor.
org.apache.flink.table.legacy.descriptors.Schema	This class was used for legacy connectors using Descriptor.
org.apache.flink.table.legacy.sources.RowtimeAttributeDescriptor	This interface will not be supported in the new source design around DynamicTableSource. Use the concept of computed columns instead. See FLIP-95 for more information.
org.apache.flink.table.legacy.sources.tsextractors.TimestampExtractor	This interface will not be supported in the new source design around DynamicTableSource. Use the concept of computed columns instead. See FLIP-95 for more information.
org.apache.flink.table.legacy.types.logical.TypeInformationRawType	Use RawType instead.
org.apache.flink.table.legacy.utils.TypeStringUtils	This utility is based on TypeInformation. However, the Table & SQL API is currently updated to use DataTypes based on LogicalTypes. Use LogicalTypeParser instead.
org.apache.flink.table.sinks.TableSinkBase	This class is implementing the deprecated TableSink interface. Implement DynamicTableSink directly instead.
org.apache.flink.table.types.utils.LegacyTypeInfoDataTypeConverter	Use DataTypeFactory.createDataType(TypeInformation) instead. Note that this method will not create legacy types anymore. It fully uses the new type system available only in the planner.
org.apache.flink.table.typeutils.TimeIndicatorTypeInfo	This class will be removed in future versions as it is used for the old type system. It is recommended to use DataTypes instead. Please make sure to use either the old or the new type system consistently to avoid unintended behavior. See the website documentation for more information.
org.apache.flink.table.typeutils.TimeIntervalTypeInfo	This class will be removed in future versions as it is used for the old type system. It is recommended to use DataTypes instead. Please make sure to use either the old or the new type system consistently to avoid unintended behavior. See the website documentation for more information.

Exceptions

Exceptions	Description
org.apache.flink.table.api.AmbiguousTableFactoryException	This exception is considered internal and has been erroneously placed in the *.api package. It is replaced by AmbiguousTableFactoryException and should not be used directly anymore.
org.apache.flink.table.api.ExpressionParserException	This exception is considered internal and has been erroneously placed in the *.api package. It is replaced by ExpressionParserException and should not be used directly anymore.
org.apache.flink.table.api.NoMatchingTableFactoryException	This exception is considered internal and has been erroneously placed in the *.api package. It is replaced by NoMatchingTableFactoryException and should not be used directly anymore.

Fields

Field	Description
org.apache.flink.table.api.dataview.ListView.elementType
org.apache.flink.table.api.dataview.MapView.keyType
org.apache.flink.table.api.dataview.MapView.valueType
org.apache.flink.table.legacy.descriptors.Schema.SCHEMA_TYPE	Schema uses the legacy type key (e.g. schema.0.type = LONG) to store type information in prior v1.9. Since v1.10, Schema uses data type key (e.g. schema.0.data-type = BIGINT) to store types.
org.apache.flink.table.module.CommonModuleOptions.MODULE_TYPE	This is only required for the legacy factory stack

Methods

Method	Description
org.apache.flink.table.catalog.Catalog.getTableFactory()	Use Catalog.getFactory() for the new factory stack. The new factory stack uses the new table sources and sinks defined in FLIP-95 and a slightly different discovery mechanism.
org.apache.flink.table.catalog.Catalog.supportsManagedTable()	This method will be removed soon. Please see FLIP-346 for more details.
org.apache.flink.table.catalog.CatalogBaseTable.getSchema()	This method returns the deprecated TableSchema class. The old class was a hybrid of resolved and unresolved schema information. It has been replaced by the new Schema which is always unresolved and will be resolved by the framework later.
org.apache.flink.table.catalog.CatalogFunction.isGeneric()	There is no replacement for this function, as now functions have type inference strategies
org.apache.flink.table.catalog.CatalogTable.of(Schema, String, List<String>, Map<String, String>)	Use the builder CatalogTable.newBuilder() instead.
org.apache.flink.table.catalog.CatalogTable.toProperties()	Only a ResolvedCatalogTable is serializable to properties.
org.apache.flink.table.catalog.ResolvedCatalogBaseTable.getSchema()	This method returns the deprecated TableSchema class. The old class was a hybrid of resolved and unresolved schema information. It has been replaced by the new ResolvedSchema which is resolved by the framework and accessible via ResolvedCatalogBaseTable.getResolvedSchema().
org.apache.flink.table.connector.source.abilities.SupportsProjectionPushDown.applyProjection(int[][])	Please implement SupportsProjectionPushDown.applyProjection(int[][], DataType)
org.apache.flink.table.factories.CatalogFactory.createCatalog(String, Map<String, String>)	Use {@link this#createCatalog(Context)} instead and implement Factory instead of TableFactory.
org.apache.flink.table.factories.CatalogFactory.requiredContext()	Implement the Factory based stack instead.
org.apache.flink.table.factories.CatalogFactory.supportedProperties()	Implement the Factory based stack instead.
org.apache.flink.table.factories.FactoryUtil.createDynamicTableSink(DynamicTableSinkFactory, ObjectIdentifier, ResolvedCatalogTable, ReadableConfig, ClassLoader, boolean)	Use FactoryUtil.createDynamicTableSink(DynamicTableSinkFactory, ObjectIdentifier, ResolvedCatalogTable, Map, ReadableConfig, ClassLoader, boolean)
org.apache.flink.table.factories.FactoryUtil.createDynamicTableSource(DynamicTableSourceFactory, ObjectIdentifier, ResolvedCatalogTable, ReadableConfig, ClassLoader, boolean)	Use FactoryUtil.createDynamicTableSource(DynamicTableSourceFactory, ObjectIdentifier, ResolvedCatalogTable, Map, ReadableConfig, ClassLoader, boolean)
org.apache.flink.table.factories.FactoryUtil.createTableSink(Catalog, ObjectIdentifier, ResolvedCatalogTable, ReadableConfig, ClassLoader, boolean)	Use FactoryUtil.createDynamicTableSink(DynamicTableSinkFactory, ObjectIdentifier, ResolvedCatalogTable, Map, ReadableConfig, ClassLoader, boolean) instead.
org.apache.flink.table.factories.FactoryUtil.createTableSource(Catalog, ObjectIdentifier, ResolvedCatalogTable, ReadableConfig, ClassLoader, boolean)	Use FactoryUtil.createDynamicTableSource(DynamicTableSourceFactory, ObjectIdentifier, ResolvedCatalogTable, Map, ReadableConfig, ClassLoader, boolean) instead.
org.apache.flink.table.factories.FunctionDefinitionFactory.createFunctionDefinition(String, CatalogFunction)	Please implement FunctionDefinitionFactory.createFunctionDefinition(String, CatalogFunction, Context) instead.
org.apache.flink.table.factories.ModuleFactory.createModule(Map<String, String>)	Use ModuleFactory.createModule(Context) instead and implement Factory instead of TableFactory.
org.apache.flink.table.factories.ModuleFactory.requiredContext()	Implement the Factory based stack instead.
org.apache.flink.table.factories.ModuleFactory.supportedProperties()	Implement the Factory based stack instead.
org.apache.flink.table.functions.ImperativeAggregateFunction.getAccumulatorType()	This method uses the old type system and is based on the old reflective extraction logic. The method will be removed in future versions and is only called when using the deprecated TableEnvironment.registerFunction(...) method. The new reflective extraction logic (possibly enriched with DataTypeHint and FunctionHint) should be powerful enough to cover most use cases. For advanced users, it is possible to override UserDefinedFunction.getTypeInference(DataTypeFactory).
org.apache.flink.table.functions.ImperativeAggregateFunction.getResultType()	This method uses the old type system and is based on the old reflective extraction logic. The method will be removed in future versions and is only called when using the deprecated TableEnvironment.registerFunction(...) method. The new reflective extraction logic (possibly enriched with DataTypeHint and FunctionHint) should be powerful enough to cover most use cases. For advanced users, it is possible to override UserDefinedFunction.getTypeInference(DataTypeFactory).
org.apache.flink.table.functions.ScalarFunction.getParameterTypes(Class<?>[])	This method uses the old type system and is based on the old reflective extraction logic. The method will be removed in future versions and is only called when using the deprecated TableEnvironment.registerFunction(...) method. The new reflective extraction logic (possibly enriched with DataTypeHint and FunctionHint) should be powerful enough to cover most use cases. For advanced users, it is possible to override UserDefinedFunction.getTypeInference(DataTypeFactory).
org.apache.flink.table.functions.ScalarFunction.getResultType(Class<?>[])	This method uses the old type system and is based on the old reflective extraction logic. The method will be removed in future versions and is only called when using the deprecated TableEnvironment.registerFunction(...) method. The new reflective extraction logic (possibly enriched with DataTypeHint and FunctionHint) should be powerful enough to cover most use cases. For advanced users, it is possible to override UserDefinedFunction.getTypeInference(DataTypeFactory).
org.apache.flink.table.functions.TableFunction.getParameterTypes(Class<?>[])	This method uses the old type system and is based on the old reflective extraction logic. The method will be removed in future versions and is only called when using the deprecated TableEnvironment.registerFunction(...) method. The new reflective extraction logic (possibly enriched with DataTypeHint and FunctionHint) should be powerful enough to cover most use cases. For advanced users, it is possible to override UserDefinedFunction.getTypeInference(DataTypeFactory).
org.apache.flink.table.functions.TableFunction.getResultType()	This method uses the old type system and is based on the old reflective extraction logic. The method will be removed in future versions and is only called when using the deprecated TableEnvironment.registerFunction(...) method. The new reflective extraction logic (possibly enriched with DataTypeHint and FunctionHint) should be powerful enough to cover most use cases. For advanced users, it is possible to override UserDefinedFunction.getTypeInference(DataTypeFactory).
org.apache.flink.table.legacy.api.TableColumn.of(String, DataType)	Use TableColumn.physical(String, DataType) instead.
org.apache.flink.table.legacy.api.TableSchema.Builder.field(String, TypeInformation<?>)	This method will be removed in future versions as it uses the old type system. It is recommended to use TableSchema.Builder.field(String, DataType) instead which uses the new type system based on DataTypes. Please make sure to use either the old or the new type system consistently to avoid unintended behavior. See the website documentation for more information.
org.apache.flink.table.legacy.api.TableSchema.fromTypeInfo(TypeInformation<?>)	This method will be removed soon. Use DataTypes to declare types.
org.apache.flink.table.legacy.api.TableSchema.getFieldType(int)	This method will be removed in future versions as it uses the old type system. It is recommended to use TableSchema.getFieldDataType(int) instead which uses the new type system based on DataTypes. Please make sure to use either the old or the new type system consistently to avoid unintended behavior. See the website documentation for more information.
org.apache.flink.table.legacy.api.TableSchema.getFieldTypes()	This method will be removed in future versions as it uses the old type system. It is recommended to use TableSchema.getFieldDataTypes() instead which uses the new type system based on DataTypes. Please make sure to use either the old or the new type system consistently to avoid unintended behavior. See the website documentation for more information.
org.apache.flink.table.legacy.api.TableSchema.toRowType()	Use TableSchema.toRowDataType() instead.
org.apache.flink.table.legacy.descriptors.Schema.field(String, TypeInformation<?>)	This method will be removed in future versions as it uses the old type system. Please use Schema.field(String, DataType) instead.
org.apache.flink.table.legacy.factories.TableSinkFactory.createTableSink(Map<String, String>)	TableSinkFactory.Context contains more information, and already contains table schema too. Please use TableSinkFactory.createTableSink(Context) instead.
org.apache.flink.table.legacy.factories.TableSourceFactory.createTableSource(Map<String, String>)	TableSourceFactory.Context contains more information, and already contains table schema too. Please use TableSourceFactory.createTableSource(Context) instead.
org.apache.flink.table.legacy.sinks.TableSink.configure(String[], TypeInformation<?>[])	This method will be dropped in future versions. It is recommended to pass a static schema when instantiating the sink instead.
org.apache.flink.table.legacy.sinks.TableSink.getFieldNames()	Use the field names of TableSink.getTableSchema() instead.
org.apache.flink.table.legacy.sinks.TableSink.getFieldTypes()	Use the field types of TableSink.getTableSchema() instead.
org.apache.flink.table.legacy.sinks.TableSink.getOutputType()	This method will be removed in future versions as it uses the old type system. It is recommended to use TableSink.getConsumedDataType() instead which uses the new type system based on DataTypes. Please make sure to use either the old or the new type system consistently to avoid unintended behavior. See the website documentation for more information.
org.apache.flink.table.legacy.sources.TableSource.getReturnType()	This method will be removed in future versions as it uses the old type system. It is recommended to use TableSource.getProducedDataType() instead which uses the new type system based on DataTypes. Please make sure to use either the old or the new type system consistently to avoid unintended behavior. See the website documentation for more information.
org.apache.flink.table.legacy.sources.TableSource.getTableSchema()	Table schema is a logical description of a table and should not be part of the physical TableSource. Define schema when registering a Table either in DDL or in TableEnvironment#connect(...).
org.apache.flink.table.plan.stats.ColumnStats.getMaxValue()
org.apache.flink.table.plan.stats.ColumnStats.getMinValue()
org.apache.flink.table.types.logical.utils.LogicalTypeParser.parse(String)	You should use LogicalTypeParser.parse(String, ClassLoader) to correctly load user types
org.apache.flink.table.types.utils.DataTypeUtils.projectRow(DataType, int[][])	Use the Projection type
org.apache.flink.table.types.utils.TypeConversions.fromDataTypeToLegacyInfo(DataType)	Please don't use this method anymore. It will be removed soon and we should not make the removal more painful. Sources and sinks should use the method available in context to convert, within the planner you should use either InternalTypeInfo or ExternalTypeInfo depending on the use case.
org.apache.flink.table.types.utils.TypeConversions.fromLegacyInfoToDataType(TypeInformation<?>)	Please don't use this method anymore. It will be removed soon and we should not make the removal more painful. Sources and sinks should use the method available in context to convert, within the planner you should use either InternalTypeInfo or ExternalTypeInfo depending on the use case.
org.apache.flink.table.utils.EncodingUtils.loadClass(String)	Use EncodingUtils.loadClass(String, ClassLoader) instead, in order to explicitly provide the correct classloader.

Constructors

Constructor	Description
org.apache.flink.table.api.dataview.ListView(TypeInformation<?>)	This method uses the old type system. Please use a DataTypeHint instead if the reflective type extraction is not successful.
org.apache.flink.table.api.dataview.MapView(TypeInformation<?>, TypeInformation<?>)	This method uses the old type system. Please use a DataTypeHint instead if the reflective type extraction is not successful.
org.apache.flink.table.expressions.CallExpression(FunctionIdentifier, FunctionDefinition, List<ResolvedExpression>, DataType)	Use CallExpression.permanent(FunctionIdentifier, FunctionDefinition, List, DataType) or CallExpression.temporary(FunctionIdentifier, FunctionDefinition, List, DataType) instead.
org.apache.flink.table.legacy.api.TableSchema(String[], TypeInformation<?>[])	Use the TableSchema.Builder instead.
org.apache.flink.table.plan.stats.ColumnStats(Long, Long, Double, Integer, Number, Number)
org.apache.flink.table.types.logical.SymbolType(boolean, Class<T>)	Symbol types have been simplified to not require a class.