Closed moshthepitt closed 6 years ago
Would we store this in a table where each row is a date and each column is the to-date aggregation?
Right now the to-date aggregate numbers we need are:
I'll confirm whether there aren't any more that we'll need
@claycrosby I think we could store all the aggregations we need for a date, including any running totals
OK cool, so just to be clear, we're now looking at 3 ETLs total for ZEIR:
1) registrations (ticket #10) 2) encounters (ticket #11) 3) to-date aggregations (this ticket)
Each of these ETLs would correspond to its own table in Superset
I'll update this ticket with any other aggregations we need beyond the 3 listed in my post above
Closing because was able to accomplish this using some nested SQL statements for ZEIR Dropout metrics (tickets #17 and #18 )
Create a new CSV output which has data aggregated by date for all the aggregations that we typically need.
This would make consumption of the output data much more easier and pleasant.
@claycrosby @jasonrogena @keymane @githengi