Does specifying interval 'DAY' instead of 'MONTH' for monthly measures avoid any aggregation treatment?


#1

So I have monthly measures, of this kind: (billID, start, end) knowing that the correct business specification sets the difference between start, end of one month. So I assume I receive data as so.
Let’s assume in rare cases, I can receive other intervals.

In my case, I don’t want any treatment, except de-duplication and interval overlapping. Setting interval to day, and treatment to integral, seems to be the only solution for me. Am I correct?

Thanks in advance.


#2

if your data interval is MONTH then it’s better to use month on your series/metric interval, if you put day then there will be disaggregation, i.e. many (30) points per month interval.


#3

Beside performance resulting of de-aggregation, is it like I get my raw data saved ?
The problem with monthly interval, is that of course my start, end, measures are not aligned with start, end of annual months, thus, C3 will do a defined treatment to aggregate on annual months. in my understanding.


#4

In that case yes i recommend using DAY interval. Look forward to improvements in v8 where you can do an evaluation like:

start: Jan5, end: Feb5, interval: 31d

Which is exactly what you’re looking for.

Also consider a query like the following:

expression: "aggregate('SUM', YOUR_METRIC)", start: Jan5, end: Feb5, interval: DAY

This should give you a single value represnting the sum of the values for each day between start and end