diff --git a/scripts/build_industrial_distribution_key.py b/scripts/build_industrial_distribution_key.py index fe7f5d82..3c62ca1b 100644 --- a/scripts/build_industrial_distribution_key.py +++ b/scripts/build_industrial_distribution_key.py @@ -96,12 +96,15 @@ def prepare_hotmaps_database(regions): # the .sjoin can lead to duplicates if a geom is in two regions if gdf.index.duplicated().any(): import pycountry + # get all duplicated entries duplicated_i = gdf.index[gdf.index.duplicated()] # convert from raw data country name to iso-2-code - s = df.loc[duplicated_i, "Country"].apply(lambda x: pycountry.countries.lookup(x).alpha_2) + s = df.loc[duplicated_i, "Country"].apply( + lambda x: pycountry.countries.lookup(x).alpha_2 + ) # Get a boolean mask where gdf's country column matches s's values for the same index - mask = gdf['country'] == gdf.index.map(s) + mask = gdf["country"] == gdf.index.map(s) # Filter gdf using the mask gdf_filtered = gdf[mask] # concat not duplicated and filtered gdf