All Products
Search
Document Center

MaxCompute:Compatibility between components and data types

Last Updated:Mar 26, 2026

MaxCompute integrates with multiple Alibaba Cloud components. Data type edition support varies by component — not all components support the MaxCompute V2.0 data type edition, complex data types, or the Hive-compatible data type edition. Check the table below before using a component with a specific data type edition.

ComponentMaxCompute V1.0 data type editionMaxCompute V2.0 data type editionComplex data typeHive-compatible data type editionDECIMAL type in MaxCompute V2.0 data type edition
StorageSupportedSupportedSupportedSupportedSupported
SDKSupportedSupportedSupportedSupportedSupported
SQLSupportedSupportedSupportedSupportedSupported
JDBCSupportedSupportedSupportedSupportedSupported
TunnelSupportedSupportedSupportedSupportedSupported
DataWorks (Data Integration)SupportedSupportedSupportedSupportedSupported
GraphSupportedSupportedSupportedSupportedNot supported
DataXSupportedSupportedSupportedSupportedNot supported
DataWorks (Data Quality)SupportedSupportedSupportedSupportedNot supported
Machine Learning Platform for AI (PAI)Supported (except DECIMAL)Supported (INT type only)Partially supported (MAP type only)Not supportedNot supported
PyODPSSupportedPartially supported ¹SupportedPartially supported ²Not supported
MapReduceSupportedNot supportedNot supportedNot supportedNot supported
OpenMapReduceSupportedNot supportedNot supportedNot supportedNot supported
DataHubSupportedNot supportedNot supportedNot supportedNot supported
LightningSupportedNot supportedNot supportedNot supportedNot supported

¹ PyODPS — MaxCompute V2.0 data type edition: DataFrame does not support this edition. Other PyODPS features support it.

² PyODPS — Hive-compatible data type edition: DataFrame does not support this edition. Other PyODPS features do not support the DATE type of the MaxCompute V2.0 data type edition.

Limitations and migration guidance

V1.0-only components: MapReduce, OpenMapReduce, DataHub, and Lightning support only the MaxCompute V1.0 data type edition. The V2.0 data type edition, complex data types, and the Hive-compatible data type edition are not supported. If you need V2.0 data type support, use alternative compute or integration components such as SQL, JDBC, or DataWorks (Data Integration).

PAI: The V1.0 data type edition is supported except for DECIMAL. In the V2.0 edition, only INT is supported among the newly added data types. Of the complex data types, only MAP is partially supported.

PyODPS DataFrame: The MaxCompute V2.0 and Hive-compatible data type editions are not supported. Other PyODPS features support the V2.0 edition but do not support the DATE type under the Hive-compatible edition.

DECIMAL type (V2.0): Graph, DataX, and DataWorks (Data Quality) do not support the DECIMAL type in the MaxCompute V2.0 data type edition.