ODBC stages are used to allow Datastage to connect to any data source that represents the Open Database Connectivity API (ODBC) standard.
ODBC stages are mainly used to extract or load the data. However, ODBC stage may also be very helpful when aggregating data and as a lookup stage (in that case it can play role of aggregator stage or a hash file and can be used instead).
Each ODBC stage can have any number of inputs or outputs.
If a processing target is an Oracle database, it may be worth considering use of ORACLE (ORAOCI9) stage. It has a significantly better performance than ODBC stage and allows setting up more configuration options and parameters native to the Oracle database.
Theres a very useful option to issue an SQL before or after main data flow operations (Oracle stage properties -> Input -> SQL). For example, when loading a big chunk of data into an oracle table, it may increase performance to drop indexes in a before SQL tab and create indexes and analyze table in a after SQL tab ('ANALYZE TABLE xxx COMPUTE STATISTICS' SQL statement).