Dropping a column in a geodataframe returns a regular dataframe

Provided a geopandas’ geodataframe, whenever I drop a column like I do with a regular pandas dataframe (see code below), it gets transformed into a regular dataframe, and hence, losing its geospatial properties. Is there any way to prevent that?

In [23]:type(gdf)
Out [23]: geopandas.geodataframe.GeoDataFrame

In [24]: gdf = gdf.drop(columns=['tmp_category'])

In [25]: type(gdf)
Out [26]: pandas.core.frame.DataFrame

[ccamara]

Thank you for your explanations and sorry of the inconvenience. It is good to know about geopandas behaviour (it may seem obvious for you, but it wasn’t for me due to the fact that is not what I have been experiencing in R) as well as how you prefer to deal with issues and questions (which largely differ from project to project).

Oh, and by the way, thanks for developing and maintaining (I’ve been greatly surprised on how fast and nicely you have dealt with this issue) geopandas, which I realise I will be using on a daily basis for my python projects.

[martinfleis]

GeoDataFrame stores geometry within the same DataFrame as every other data. GeoDataFrame without geometry column is not considered as valid, hence it may be returned as DataFrame in some occasions. I believe that resolves the issue, so I am closing it now. Next time there’s no need to open an issue, we are monitoring StackOverflow, so I also seen your question there as well. SO is a better place for this kind of debugging.