| 摘要 |
<p>A computer implemented method for processing data on a distributed database management system implemented on a network of multiple nodes (115), each of the multiple nodes (115) comprising at least one of a processor (117), storage device (121), and memory (119). The method comprises receiving a query referencing first and second user-defined table functions and a table (T) distributed on the multiple nodes (115); executing the first user-defined table function on each of the multiple nodes (115) by processing records of the table (T) residing on the respective nodes (115A,115B) for generating a data sample per node; storing the generated data samples in a selected, preferably by the user in the query, set of nodes (115D); executing the second user-defined table function on each of the selected set of nodes by processing records of the data samples residing on the respective nodes of the selected set of nodes to compute a user-defined aggregate.</p> |
