We are going to move our time series data into timestream, but I can not find any information about how to do that, our IOT data come to s3 as a million small files(JSON) on a daily basis, and each JSON looks like this:
{'meter_id':'123', 'A':[{ts:x,value:xx},{},...48 (obects)], 'B':[{},{},...48 (obects)]}
In the real world,'A' and 'B' are in a deeper level.
Those 48 objects includes timeseries(30min interval timeseries) and reading data, how to do something like posexplore in pyspark then put those data into Timestream, seems that all the official examples are using flatten dataset.
