Spring integration: use Splitter and Aggregator

Question

I'm reading files line by line using FileSplitter, and then filter some of the corrupted lines - as to discard some lines - in order to do webservices calls for each line.

What I need to do in addition to this is to get the file moved to completed files directory.

Update: Just to be more clear, the successful/good lines after filtering will need to be aggregated in a file to be written to the completed directory and the corrupted/filtered lines will be written to another one.

Mentioning the above, I was not clear about the following points:

• Will I need some kind of correlation between the messages/lines as to provide it manually or Spring Integration will do it for me?
• If some lines are discarded due to the filtering, what is the right strategy to use with the aggregator?
• How can I make use of file markers as to identify when the file is completed to move to the completed directory?

The scenario simply looks like:

FileSplitter -> line Filter -> Outbound Gateway (webservices) -> Aggregator

Answer 1

The FileSplitter can be configured to emit FileSplitter.FileMarker as the first message with the FileSplitter.FileMarker.Mark.START and in the end with the FileSplitter.FileMarker.Mark.END respectively.

You can add a router (PayloadTypeRouter should be enough) after splitter to handle FileSplitter.FileMarker in a different flow. And if it is a FileSplitter.FileMarker.Mark.END you just perform a desired logic to move to the completed directory.

I don't think that you need an aggregator for this scenario.

Anyway the FileSplitter can be configured with the setApplySequence(true) to be able to correlate a group downstream in the aggregator. But you still won't be able to know the size of the group to release it in time. The FileSplitter.FileMarker.Mark.END still can help here, although I don't see the reason in that for your task.

There is some info on the matter in the Reference Manual.