process() method

public int process()

The process() method is the entry point for processing input and output rows. Override the process() method in your Stage subclass.

The Java Pack API calls the process() method every time an event occurs:

A single process() call can handle the complete set of input records. However, if you want to monitor, through InfoSphere® DataStage® Director, the number of records that are processed within a timeframe, it is recommended that process() returns frequently.

Reading rows

In your process() implementation, you can call the readRow() method to read input rows. The effect of not calling the readRow() method before returning from the process() method depends on which Java Pack stage you use and how it is deployed in your job.
Stage / Deployment Effect of Omitting a readRow()
Java Transformer InfoSphere DataStage stops the job and writes the following message to the job log:
Deadlock detected: all input links
blocked
Java Client (target) The input row is discarded.
Java Client (lookup)

The active stage that is connected to the Java Client stage uses any rows that the Java Client writes on the link.

Default Implementation

The default process() implementation reads the next available input row and writes its contents to the job log. These log entries are created only when tracing has been activated for the job through InfoSphere DataStage Director.

Returns

The process() method must return one or more of these status fields: