If you draw a link from your Oracle stage and change the Oracle stage settings then all rejects will go down that link.
There are so many factors affecting throughput that the second question is not easily answered, it is like asking "how can I make my car faster?" and expecting a concise response.
Usually most of the factors affecting throughput are changed on the database side, not just the DataStage side.
I could try to explain, but could not come close to the numerous detailed descriptions in the Parallel Job Developer's Guide. The documentation details reject handling for all stages that support this output.
Hi Harsha,
You can capture the reject records while you use the oracle load method. The rejected records while loading the data into the table gets saved in the scratch disk with the control file,log file and thebad file. The bad file is the bad data which failed to load the table.
Once you use the load method, the job aborts if the data is rejected,to catch all the rejected records then you need to use the APT_ORACLE_LOAD_OPTIONS properly to get all rejected records. Once the job completed(aborted), you should get the control,log and bad file from the scratch and place it into the different path.