Open pkdash opened 1 year ago
For timeseries aggregation, data for a given series is loaded into pandas DataFrame. This data is loaded from a single table - TimeSeriesResultValues only. That means the user will be able to edit data for this only one table. Not sure if this is adequate for editing timeseries aggregation data.
Editing aggregation data for raster, netcdf and geo-feature using the modules specified above may not be adequate. User may have to depend on additional packages to meet their editing requirements for these aggregation types. Primarily the module we use here to load the data for any of these aggregation types, would allow the user to read the aggregation data which then can be used to create new data files (most likely using other python packages) for that aggregation. These new updated data files can then be used to update an aggregation or saved as a new aggregation using hsclient.
Here is the list of aggregations and the corresponding data processing objects:
When the underlying data is modified using the data processing object, the user should be able to save the modified data as a new aggregation or update the data of the current aggregation. It's expected that the user is familiar with the above listed data processing python modules and can use these modules to create/update data for the relevant aggregation types.