Closed FelixMau closed 1 year ago
In 0baca0e76669efafd004e850ef8a511594f32ffc I added "year" to scalars. The region
is inherently mapped by the profile
column since each region should have its own profile. However, I can imagine a scenario where future users may only want scalar data to vary across regions.
Timeseries are currently only matched if field type is matching as a Series
and if the Series entries are all found within timeseries columns. This approach assures to only write foreign keys for timeseries if there are Series connected to the Dataset with matching naming.
Timeseries are exploded within data_adapter
therefore this issue is outdated and the period
csv writing will be tackled in #44
There has to be additional functionality to write correct time series names in columns for time series. The current situation is like this:
initial column name
(data) +region
e.g. _onshoreHHinitial column name
(facade) +region
e.g _profileHHAlthough using a similar method, the result is different as different datasets/columns are referenced. This must somehow be unified:
GLOBAL_PARAMETER_MAP.yaml
in case there are multiple time series for one processThere might be cases where more time series are allowed in the facade as there are provided time series.
Further:
process_data.scalar
has a row for eachregion
andyear
, the refactored time series only have one column for eachregion
and allyears
. This is not a big problem, but needs to be addressed at some point when it will be integrated into tabular.