All filtering examples in apache flink documentation display simple cases of filtering according to a global threshold.
But what if filtering on an entry should take into account the previous entry?
Let's say we have a stream of sensor data. We need to discard the current sensor data entry if it's X% larger than then previous entry.
Is there a simple solution for this? Either in Apache Flink or in plain Java.
Thanks
In flink, this can be done with state.
Your use case is very similar to the fraud detection example from flink doc.