I realize the program could be a web, FTP, etc. server, but I'm talking about directly & efficiently. Having to start a program over & over again seems a huge waste when throughput goes up. Imagine calling a running Java program that is in a while() loop, looping through on each line from STDIN, using a newline as the delimiter. ExecuteStreamCommand seems to just call a program/script, optionally passing the flowfile on STDIN, but again, it is asking to load the program from disk & GC for every single flowfile... Can we send data to that program lightning fast without having to setup web, SFTP, SSH servers? --- Again, the use case is a processor being hammered with text as a flowfile, and wanting to get it into a java program as fast as possible. Loading, running, and stopping the code for each and every flowfile is just too computationally expensive. Preferred is a direct connection to the application, as when the application is running continuously, blocking when no data. Your insights appreciated! --- We're not talking about calling a static method, but this at first seemed interesting: from here: "Answer by mburgess Apr 27 at 03:34 AM To @Artem Ervits comment, if the Java class is available in the class loader (either by being part of the NiFi Core or being specified as a JAR in the ExecuteScript's Module Directory property), then you can call it directly, you don't need reflection. If I'm not answering your question, can you provide more details? I'm happy to help :)"
... View more
