尝试将多边形中的两个多边形元素分配给 geopandas 几何列会导致 ValueError

Question

我正在使用世界地图中内置的 geopandas。我正在尝试将法属圭亚那从法国几何中分离出来，并为法属圭亚那创建一个新条目（我已经成功完成）。但是，当将缩减的欧洲法国和科西嘉岛的多面体重新评估回法国 geometry 单元格时，出现错误。

world = gpd.read_file(gpd.datasets.get_path('naturalearth_lowres'))

# Remove French Guiana (shape[0])
shape = world[world['name'] == 'France']['geometry'].all()

fr_shape = shape[2] # This works creating a POLYGON but drops Corsica :( 
world.at[world['name'] == 'France', 'geometry'] = fr_shape

fr_shape = shape[1:] # This creates a MULTIPOLYGON then throws an ValueError.
world.at[world['name'] == 'France', 'geometry'] = fr_shape

> ValueError: Must have equal len keys and value when setting with an iterable

与此处类似的问题：Geopandas set geometry: ValueError for MultiPolygon "equal len keys and value"

但是，当我试图提取多面体的三个元素中的两个并重新分配时，对我来说似乎是一个不同的问题，因为另一个是从一个数据帧直接复制到另一个数据帧而无需操作。迄今为止，尝试.values 解决方案的各种变体并没有成功抛出相同的问题。

Answer 1

我已经设法找到解决方法：

import pandas as pd
import geopandas as gpd

world = gpd.read_file(gpd.datasets.get_path('naturalearth_lowres'))

# Remove French Guiana from France.
shape = world[world['name'] == 'France']['geometry'].all()

# Multipolygon ValueError Workaround.
fr_df = pd.Series(['France', 'France'], name='country')
fr_df = gpd.GeoDataFrame(fr_df, geometry=[shape[1], shape[2]])
fr_df = fr_df.dissolve(by='country')
world.at[world['name'] == 'France', 'geometry'] = fr_df['geometry'].values                     

world.plot()