Hi, I am newbie in Flink and have questions about stream states. I can’t find answers in documentation, but if I just miss one, please link to doc)
1. Is ValueState (and other state classes) are ‘stream' scoped? So that, it is not possible to share same state with two (or more) different pipelines in a same job: - sourceA -> keyBy -> mapWithStateXUpdate -> print - sourceB -> keyBy -> mapUsingStateXValue -> print Operators ‘mapWithStateXUpdate’ and ‘mapUsingStateXValue’ will use separate copy of StateX even they use same name for it and same key (ValueStateDescriptor(“StateX”…)). Is it right? 2. In Streaming Guide about connect() operation: == Connect allowing for shared state between the two streams. == But how do access state from operators after connect - connect() returns ConnectedStream but not Keyed, so states defined early is not accessible. Or it means that doing keyBy after connect() will allow to define new state based on values from both streams? 3. Is state are ‘operator’ scoped? - source -> KeyBy -> mapWithStateX_1 -> keyBy -> mapWithStateX_2 Assume both map try to use state with same name (ValueStateDescriptor(“StateX”…)). But despite of that, they will have different copies. is it right? Thanks! -- Andrey Utkin <cind...@gmail.com <mailto:cind...@gmail.com>>
