| 摘要 |
Reroutable data columns are identified in an ETL process by receiving an ETL process definition describing a set of processing stages and how each processing stage output data column is a result of a function that operates on a set of input data columns, representing the ETL process definition as a directed graph with nodes representing processing stages and links representing data flow between processing stages, traversing at least part of the directed graph and identifying a set of subsequent nodes of the directed graph where at least one data column is involved only as input data in identity functions, the at least one data column being reroutable between outmost nodes of the set of subsequent nodes, and in connection with traversing the at least part of the directed graph, maintaining information about reroutable data columns and the respective outmost nodes. |
