This section discusses the limitations and known issues in the MySQL NDB Cluster Connector for Java APIs.
Joins: With ClusterJ, queries are limited to single tables. This is not a problem with JPA or JDBC, both of which support joins.
Database views: Because MySQL database views do not use the
NDBstorage engine, ClusterJ applications cannot “see” views, and thus cannot access them. To work with views using Java, you should use JPA or JDBC.
Relations and inheritance: ClusterJ does not support relations or inheritance. Tables are mapped one-to-one onto domain classes, and only single-table operations are supported.
NDBtables for NDB Cluster 7.3 and later support foreign keys, and foreign key constraints are enforced when using ClusterJ for inserts, updates, and deletes.
TIMESTAMP: Currently, ClusterJ does not support the TIMESTAMP data type for a primary key field.
Known issues in JDBC and Connector/J: For information about limitations and known issues with JDBC and Connector/J, see JDBC API Implementation Notes, and Troubleshooting Connector/J Applications.
Known issues in NDB Cluster: For information about limitations and other known issues with NDB Cluster, see Known Limitations of NDB Cluster.