When you are creating a Java transformation and go to the Java Code tab to fill out the custom code you want to be executed at run-time, there are three main sections that are focused on and widely used:
The main purpose and the difference between the three is as follows:
The code in this section gets executed when a row is passed to the Java transformation. For example, in this section you can put code to assign values from input ports to output ports (including the manipulation you want to do to the incoming data).
The code in this section gets executed on end of data (for every partition in case you have multiple partitions). For example, this section can be used to include code that will print out statistical kind of data (like how many rows were written, what was the highest number, etc).
This section is used to create methods that will get called in either one of the two sections above. In the expressions you can call on lookups and do anything that an Expression transformation would do.
What can we do to improve this information (2000 or fewer characters)