GeoPandas to_crs方法报错ProjError：x,y,z,time维度不一致求助

问题概述 #

使用GeoPandas的to_crs方法转换坐标参考系时触发错误：ProjError: x, y, z, and time must be same size，已执行conda update --all并在多环境测试，问题未解决。

可复现代码 #

### to crs method check
import geopandas as gpd
import pandas as pd
from shapely.geometry import Point

df = pd.DataFrame(
    {'City': ['Buenos Aires', 'Brasilia', 'Santiago', 'Bogota', 'Caracas'],
     'Country': ['Argentina', 'Brazil', 'Chile', 'Colombia', 'Venezuela'],
     'Latitude': [-34.58, -15.78, -33.45, 4.60, 10.48],
     'Longitude': [-58.66, -47.91, -70.66, -74.08, -66.86]})

gdf = gpd.GeoDataFrame(df, geometry=gpd.points_from_xy(df.Longitude, df.Latitude))
gdf.crs = "EPSG:4326" # original crs
gdf = gdf.to_crs("EPSG:6668") # convert to new crs
print(gdf)

报错栈 #

ProjError                                 Traceback (most recent call last)
Cell In[18], line 14
     12 gdf = gpd.GeoDataFrame(df, geometry=gpd.points_from_xy(df.Longitude, df.Latitude))
     13 gdf.crs = "EPSG:4326" # original crs
---> 14 gdf = gdf.to_crs("EPSG:6668") # convert to new crs
     15 print(gdf)

File ~\anaconda3\lib\site-packages\geopandas\geodataframe.py:1364, in GeoDataFrame.to_crs(self, crs, epsg, inplace)
   1362 else:
   1363     df = self.copy()
-> 1364 geom = df.geometry.to_crs(crs=crs, epsg=epsg)
   1365 df.geometry = geom
   1366 if not inplace:

File ~\anaconda3\lib\site-packages\geopandas\geoseries.py:1124, in GeoSeries.to_crs(self, crs, epsg)
   1047 def to_crs(self, crs=None, epsg=None):
   1048     """Returns a ``GeoSeries`` with all geometries transformed to a new
   1049     coordinate reference system.
   1050 
   (...)
   1121 
   1122     """
   1123     return GeoSeries(
-> 1124         self.values.to_crs(crs=crs, epsg=epsg), index=self.index, name=self.name
   1125     )

File ~\anaconda3\lib\site-packages\geopandas\array.py:779, in GeometryArray.to_crs(self, crs, epsg)
    775     return self
    777 transformer = Transformer.from_crs(self.crs, crs, always_xy=True)
-> 779 new_data = vectorized.transform(self.data, transformer.transform)
    780 return GeometryArray(new_data, crs=crs)

File ~\anaconda3\lib\site-packages\geopandas\_vectorized.py:1114, in transform(data, func)
   1111 result[~has_z] = set_coordinates(data[~has_z].copy(), np.array(new_coords_z).T)
   1113 coords_z = get_coordinates(data[has_z], include_z=True)
-> 1114 new_coords_z = func(coords_z[:, 0], coords_z[:, 1], coords_z[:, 2])
   1115 result[has_z] = set_coordinates(data[has_z].copy(), np.array(new_coords_z).T)
   1117 return result

File ~\anaconda3\lib\site-packages\pyproj\transformer.py:430, in Transformer.transform(self, xx, yy, zz, tt, radians, errcheck, direction)
    428     intime = None
    429 # call pj_transform.  inx,iny,inz buffers modified in place.
-> 430 self._transformer._transform(
    431     inx,
    432     iny,
    433     inz=inz,
    434     intime=intime,
    435     direction=direction,
    436     radians=radians,
    437     errcheck=errcheck,
    438 )
    439 # if inputs were lists, tuples or floats, convert back.
    440 outx = _convertback(xisfloat, xislist, xistuple, inx)

File pyproj/_transformer.pyx:459, in pyproj._transformer._Transformer._transform()

ProjError: x, y, z, and time must be same size

环境信息 #

• 系统：Windows 10 10.0.22621-SP0，Python 3.9.16
• 核心库版本：
  • GEOS: 3.8.0
  • GDAL: 3.6.2
  • PROJ: 6.2.1
  • geopandas: 0.12.2
  • pyproj: 2.6.1.post1
  • shapely: 2.0.1

解决思路与方案 #

1. 版本兼容性修复（优先推荐） #

错误根源是Shapely 2.0.x与旧版pyproj/PROJ的兼容性冲突：Shapely 2.0对几何坐标的Z值处理逻辑更新，而pyproj 2.6.1、PROJ 6.2.1不支持新逻辑，导致转换时Z坐标维度匹配失败。

执行以下conda命令升级到兼容版本：

conda install -c conda-forge pyproj>=3.0 proj>=7.0 geopandas shapely

该命令会安装与Shapely 2.0匹配的pyproj/PROJ版本，同时确保GeoPandas依赖兼容。

2. 强制转换为2D几何 #

如果暂时无法升级库，可以强制将所有几何对象转为2D，避免Z坐标干扰转换流程：
修改代码，在设置CRS后添加以下行：

# 转换所有几何为2D
gdf['geometry'] = gdf['geometry'].apply(lambda geom: geom.to_2d() if hasattr(geom, 'to_2d') else geom)

之后再执行to_crs转换即可。

3. 验证坐标系定义 #

若上述方案无效，可先尝试转换到常见坐标系（如EPSG:3857）：

gdf = gdf.to_crs("EPSG:3857")

如果转换成功，说明问题出在EPSG:6668的坐标系定义——旧版PROJ的坐标参考系数据可能缺失或不兼容该EPSG代码，此时需升级PROJ到7.0+版本（方案1）。

内容的提问来源于stack exchange，提问作者wayway0914