AnsweredAssumed Answered

What is the best way to store a Blob of (time,value) pair forecasts for historical purposes?

Question asked by jagdish.konathala Champion on Sep 20, 2017
Latest reply on Feb 21, 2019 by jagdish.konathala

Use case : Future, predicted forecast data to be stored that is received every 5 minutes. So, every 5 minutes, we get a 24 hours forecast that needs to be overwritten as time progresses. For historical forecast analysis, we want to keep track of this 8 hours worth of events (96 events - 5 minute spaced).


Here is the design : We have two sets of points, one set with data that is continuously overwritten as it changes. The other set that should Blob the entire time value pair as one event {(timestamp1, value1), (timestamp2, value2), (timestamp3, value3), ...(timestamp96, value96)}. Blob can only 967 bytes and my sample is about 3KB.


PI Web API is what is used to write this forecast data in the 2 sets of points (realtime and historical blobs).


Any suggestions on a better, more efficient way of achieving this?