Take the following data series from a volume metet's accumulated volume for example:
|day| reading |
|d1, |1|
|d2, |2|
|d3, |missing (3)|
|d4, |missing (4)|
|d5, |missing (5)|
|d6, |6|
|d7, |7|
If a query needs to show daily usage, EXCLUDING the accumulated error due to delay in data collection, communication error etc. I.e. for day d6, the result should be null, instead of 4. I am not sure how to filter this 4. I can do a post median aggregation but that will skew all other values.
Is there any way to ensure all empty data points to generate empty delta after difference transformation