I need to run unit test case inputs at high speed (many test cases) in regression test that will preserve relative order with multiple input streams.

Run unit test case inputs at high speed in regression test that will preserve relative order with multiple input streams.

This is difficult because even when the data is in the correct order on the input queues for each of the input streams, which input stream is processed next is non-deterministic and if there is a data-skip where one queue "should" be read twice in a row to preserve order, there is no way to tell the engine that, hence the documented limitation.

Either remove the requirement that order be preserved across multiple input streams, or change the test configuration so that only one input stream is needed.

To convert to use a single input stream, combine all input-stream schemas into one and merge the CSV input data similarly. Then in the app use a Filter and Maps to determine where each message should be routed and use Synchronous container connections to connect the "mapper" output streams to the application input streams.