dhis2-devs team mailing list archive

[Jason Pickering <jason.p.pickering@xxxxxxxxx>]

>
>
> OK. That's where the mentioned system setting comes in ("Omit indicator
> values with zero numerator value"). The difference between the 0s in the
> agg.datavalue table and the 0s you mention in the agg.indicatorvalue is that
> the lattter has a non-0 denominator value. (Entries with 0 denominator
> values are skipped completely). These values might be interesting when you
> are working eg. with pivot tables, which will use both the numerator and the
> denominator value when performing aggregation on its own. This is the reason
> why we included both numerator and denominator values in the
> agg.indicatorvalue table in the first place.

OK, but then I would think that zeros should also be written to the
aggregatedatavalue table as well ?

If the aggregated numerator shows up as the numerator in the
aggregatedindicatorvalue table as zero (but not blank) but does not
show up at all in the aggregateddatavalue table at all (neither as
blank or zero) it does not seem consistent to me.

Shouldn't  the result of an aggregation operation of all blanks
results in a zero, which would be the same as an aggregation operation
of a set of all zeros or a mixture of both blanks and zeros?

 It would seem to be the case with the aggregatedindicatorvalue table,
but not the aggregateddatavalue table?


Again, looks like I need to read the documentation. ;)