There is no issue with IBM InfoSphere DataStage's graphical interface for designing data flows, but I will provide feedback that we are gathering the source from the Oracle database mainly, as well as from some spreadsheets. With respect to the Oracle DB Connector, if you write any PL/SQL or SQL with the connectors, there aren't many options, such as executing procedures in the PL/SQL, executing functions, or executing packages. The Oracle connector doesn't have many features and needs improvement. Nowadays many people are writing programs in Python or in PL/SQL with respect to Oracle, so especially in IBM InfoSphere DataStage, there are no features to call programs directly instead of calling them as a script. What I am facing, especially with parallel processing, is that a developer and admin have to sit together. They have to run the job multiple times with different combinations of parallel processing to get the best performance. Instead of that, if the job itself gave some guidance, such as running this parallel processing with this many nodes, it would help; I think that is missing. An additional feature I would want to see in the next release is the ability to work on logs, especially machine logs or artificial logs, to pull semi-structured or unstructured data without having to write extensive code in Python and integrate it. If IBM InfoSphere DataStage provided some feature for this, it would help.
