Suppose my goal is to answer the question: "Show me all the continuous time intervals a stock price was over X between times T1 and T2".
The response I am looking for is an array of Tuple2s, each of which represent a start / end timestamp when the stock was above price X. There might be a lot of these if a stock is bouncing between a tight channel.
If the stock is currently over X, the current timestamp.
I want to be able to arbitrarily set T1 and T2. These might span a year, they might span a few hours.
Is this an appropriate state /queryable state problem for Flink? Can I store this information permanently in Flink, or should I emit completed timestamps to a persistent data store, and then query both Flink and the other data store for the information?
Thanks!
Locked
