dijie

Is one Coordinator better than many? Because if an exception occurs at Sink and the Source does not know about it and continues to send data downstream, is the data...

We do translation work for different engines, so if faced with inconsistent semantics or inconsistent functions how to solve the situation? For example, Spark and Flink both have Checkpoint, but...

Anyway, looking forward to the new API . 🎉

> > We do translation work for different engines, so if faced with inconsistent semantics or inconsistent functions how to solve the situation? For example, Spark and Flink both have...

> > > We do translation work for different engines, so if faced with inconsistent semantics or inconsistent functions how to solve the situation? For example, Spark and Flink both...

> I have another question. If a Connector is not implemented in Spark or Flink, how will we implement this new Connector if we are based on a new engine?...

Thank everyone for answering my questions, maybe I need to see the code to understand the design better.

From the code, I only saw the abstract upper design about catalog,table,source,sink... and some logic of converting Seatunnel to Flink types. There is no distributed execution or network transmission or...

Or are we going to do a datax-like model for now?