search

sedos-project / data_adapter_oemof

This respository holds the data adapters to connect oemof with the OEDatamodel-concrete.
GNU Affero General Public License v3.0
0 stars 1 forks source link

Write scalars as scalars and not as lists #76

Closed nailend closed 3 months ago

nailend commented 3 months ago

In my current AP8 dataset I use, scalars are returns as list of scalar and also passed as list of scalars to tabular. which doesnt work.

this is a fix. I am not sure if this is actually necesary as @henhuy maybe already takes care of this in data_adapter

henhuy commented 3 months ago

I think they are grouped as period values, or not? thougt we would refactor them to timeseries, as tabular cannot handle period values yet?

nailend commented 3 months ago

No its about single scalars like efficiencies which are constant. These would be translated in tabular to timeseries if necessary. But currently they would end up as list of lists on tabular and not as a list of values as they are in the datapackage as a list of one single value.

FelixMau commented 3 months ago

If lists are found that do not match timeseries lengths they stay as they are because we cant actually know what those values are. This issue is couriering due to unexpected data-format in BEV data

nailend commented 3 months ago

Regardless of the origin of this problem, I am pretty certain that a list of an iterable with length 1 eg. [[1,]] or [(1,)] is for no use. In my situation, data is uploaded to oep for each year as Type: float array and not as Type: float. This is therefore, a fix for wrong datatype upload.

I think they are grouped as period values, or not? thougt we would refactor them to timeseries, as tabular cannot handle period values yet?

data-adapter-oemof does that for multiple years, but if the value for one year is a list (e.g. efficiency = [0.28]), it will be grouped as list of iterable (e.g. [[0.28,]. [0,28], ...).

I currently need this, once the uploaded data is flawless, this could be removed again.

henhuy commented 3 months ago

I think the type float array is regarding a bandwidth. But in scenario data there should not be bandwidths anymore! Thus, the APs should upload scenario data as well. In think we discussed on versioning it as s1.

nailend commented 3 months ago

I talked to Hedda. New data will take some time to be uploaded. I need this to go on with bev-development. I will therefore merge this in features/example_transport. You might revert this merge later.