Hi everyone,
first of all I want to give a quick overview of what I am currently dealing with:
I try to connect a external Python Server with my HANA in order to run some algorithms. Therefore I use ZeroMQ in combination with Protobuf and implemented these features into an adapted AFL INTRO DEMO. The reason why I am doing this is just to have a POC so we could use different kind of external servers/codes/languages without rewriting everything into AFL/C++. Also some of the code is much easier to maintain or change without rebuilding the AFL. I have to say that this is my first experience with HANA AFL SDK, so I am on my own 
Right now it seems like everything is working fine. I wrote a testscript that takes a dataset (simple double values) from an input table, sends them to the python server via afl->protobuf->zmq, receives a new dataset from the python server on the similar way and inserts new values into another table.
My Problem now is that I cannot run the script with bigger datasets of roundabout 120k values. Also repeated running of the script will lead to an error. The error for both cases is:
SAP DBTech JDBC: [139]: current operation cancelled by request and transaction rolled back: search table error: [2625] execution plan aborted
First it appeared that HANA couldn´t handle more than one request in a short period of time, so there might be some kind of caching problem, but then I retried running the script repeatedly with a very small dataset which resulted in the same problem. Also it looks like there is no problem with the python server as I can monitor the server sided process.
Maybe someone has an idea where the problem is. I really need to handle bigger datasets and repeated execution of the script.
King Regards
Kenny