handle numeric key_on values for Choropleth #1193
Conversation
This addresses a bug where geometries and data coming from GeoDataFrames were not joined. This was caused by the index being casted to string during the conversion to GeoJSON. The new behaviour is the following: - try the join as it is - if index contained numeric value, cast to str and try again This commit includes 3 new test cases where: - 1st and 2nd were failing under old behaviour - 3rd passes under both behaviours
There was a problem hiding this comment.
Hi @alessioarena,
Sorry for my late review. I've been doubting about whether we should implement this or not. My hesitation is because of the added complexity and the possible side effects I'm not seeing. Any thoughts on that?
But given that this will solve something failing in Choropleth silently I'm inclined to merge this. Could you address my 2 comments?
folium/features.py
Outdated
| # This is a pd.Series | ||
| data = data.set_index(columns[0])[columns[1]] | ||
|
|
||
| # convert the data to dictionary |
There was a problem hiding this comment.
Nit: this comment is a bit superfluous
There was a problem hiding this comment.
Noted. I can push the change, but I'm keen to wait for the other discussion to be agreed upon and resolved
There was a problem hiding this comment.
The original logic was actually not working for my test case, in which the geojson feature id is an integer and the data has a string index. I couldn't really work out what was going on, so I opted to simplify the logic instead: if the original key doesn't fit, try again with a str or int, depending on the original type. I hope this should be a bit simpler to grok.
This addresses issue #1185 where geometries and data coming from GeoDataFrames
were not joined. This was caused by the index being casted to string
during the conversion to GeoJSON.
The new behaviour is the following:
This commit includes 3 new test cases where: