nir_convert_df_to_long removes nan values #100
Labels
bug
Something isn't working
priority: high
High priority issue: prioritize and solve as soon as possible
priority: medium
medium priority issue. Good to solve soon if possible
Describe the bug
Currently the function primap2.pm2io.nir_convert_df_to_long removes values interpreted as nan. This included e.g. the string "NA". This contradicts the option to specify a mapping to nan and 0 when converting the data to primap2 format. Thus this should be changed for consistency reasons.
Expected behavior
Keep nan values, so mapping can be made explicit
It is easy to fix, just add 'dropna=False' to the stack command in line 167 of _GHG_inventory_reading.py.
Additional context
It's not worth it's own branch an pull request so the issue is just a reminder to fix this with the next set of changes.
The text was updated successfully, but these errors were encountered: