Cartopy基础示例运行失败，报ValueError错误求助

已通过conda和pip3安装Cartopy及其依赖库，运行官网基础示例代码：

ax = plt.axes(projection=ccrs.PlateCarree())
ax.coastlines()

plt.savefig('coastlines.pdf')
plt.savefig('coastlines.png')

plt.show()

触发ValueError: zero-size array to reduction operation minimum which has no identity，错误栈如下：

Traceback (most recent call last):
  File "/Users/mattbright/Desktop/SCripts/lattice_geo.py", line 6, in <module>
    ax = plt.axes(projection=ccrs.Mollweide())
  File "/Users/mattbright/opt/miniconda3/lib/python3.7/site-packages/matplotlib/pyplot.py", line 947, in axes
    return subplot(111, **kwargs)
  File "/Users/mattbright/opt/miniconda3/lib/python3.7/site-packages/matplotlib/pyplot.py", line 1125, in subplot
    ax = fig.add_subplot(*args, **kwargs)
  File "/Users/mattbright/opt/miniconda3/lib/python3.7/site-packages/matplotlib/figure.py", line 1402, in add_subplot
    ax = subplot_class_factory(projection_class)(self, *args, **kwargs)
  File "/Users/mattbright/opt/miniconda3/lib/python3.7/site-packages/matplotlib/axes/_subplots.py", line 42, in __init__
    self._axes_class.__init__(self, fig, self.figbox, **kwargs)
  File "/Users/mattbright/opt/miniconda3/lib/python3.7/site-packages/cartopy/mpl/geoaxes.py", line 355, in __init__
    super(GeoAxes, self).__init__(*args, **kwargs)
  File "/Users/mattbright/opt/miniconda3/lib/python3.7/site-packages/matplotlib/axes/_base.py", line 511, in __init__
    self.cla()
  File "/Users/mattbright/opt/miniconda3/lib/python3.7/site-packages/cartopy/mpl/geoaxes.py", line 528, in cla
    self._boundary()
  File "/Users/mattbright/opt/miniconda3/lib/python3.7/site-packages/cartopy/mpl/geoaxes.py", line 1425, in _boundary
    trans_path = proj_to_data.transform_path(path)
  File "/Users/mattbright/opt/miniconda3/lib/python3.7/site-packages/matplotlib/transforms.py", line 1527, in transform_path
    return self.transform_path_affine(self.transform_path_non_affine(path))
  File "/Users/mattbright/opt/miniconda3/lib/python3.7/site-packages/cartopy/mpl/geoaxes.py", line 181, in transform_path_non_affine
    src_path.vertices, self.source_projection)
  File "/Users/mattbright/opt/miniconda3/lib/python3.7/site-packages/cartopy/crs.py", line 649, in quick_vertices_transform
    if (x.min() >= x_limits[0] and x.max() <= x_limits[1] and
  File "/Users/mattbright/opt/miniconda3/lib/python3.7/site-packages/numpy/core/_methods.py", line 43, in _amin
    return umr_minimum(a, axis, None, out, keepdims, initial, where)
ValueError: zero-size array to reduction operation minimum which has no identity

已尝试重装Anaconda、回退Matplotlib至3.3.0，均无效，求其他解决方法。

1. 修复Cartopy自然特征数据缺失/损坏问题 #

该错误大概率是Cartopy依赖的自然地球数据（如海岸线）未正确下载或文件损坏导致的：

• 手动下载基础数据集：获取官方提供的ne_10m_coastline等基础数据，解压到Cartopy的数据目录，默认路径为：

  /Users/mattbright/opt/miniconda3/lib/python3.7/site-packages/cartopy/data/raster/natural_earth
  

• 自定义数据路径：若不想修改默认目录，可设置环境变量CARTOPY_DATA_DIR指向你的数据存储文件夹。
• 清理旧数据：删除原有损坏的数据文件后重新下载，确保文件完整性。

2. 统一依赖安装方式并指定兼容版本 #

混合使用conda和pip安装易引发依赖冲突，且版本不匹配也会导致此类错误：

• 完全用conda安装兼容版本组合：

  # 先卸载pip安装的Cartopy
  pip uninstall -y cartopy
  # 安装兼容版本的Cartopy、Matplotlib和NumPy
  conda install cartopy=0.18 matplotlib=3.3 numpy=1.19 --force-reinstall
  

  （0.18版本的Cartopy对Python3.7、Matplotlib3.3和NumPy1.19的兼容性经过验证）

3. 显式设置坐标轴范围 #

自动范围计算可能出现空数组问题，手动设置范围可规避：
修改示例代码，添加坐标轴范围设置：

import matplotlib.pyplot as plt
import cartopy.crs as ccrs

# PlateCarree投影示例
ax = plt.axes(projection=ccrs.PlateCarree())
# 显式设置经纬度范围
ax.set_extent([-180, 180, -90, 90], crs=ccrs.PlateCarree())
ax.coastlines()

plt.savefig('coastlines.pdf')
plt.savefig('coastlines.png')
plt.show()

若使用Mollweide投影，同样添加上述set_extent语句即可。

4. 调试并修复NumPy数组异常 #

错误栈指向NumPy的最小值计算，可添加调试代码确认数组状态：
在代码中插入调试语句，检查投影转换时的顶点数组：

import numpy as np
# 在创建坐标轴后添加
print("Axes boundary vertices:", ax.get_extent())
# 或者查看Cartopy内部的路径顶点
path = ax._get_boundary_path()
print("Path vertices shape:", path.vertices.shape)

若输出显示顶点数组为空，可尝试降级NumPy到1.19.x版本，避免新版本的计算逻辑冲突。

内容的提问来源于stack exchange，提问作者Matt Bright