Geopandas中Polygon对象无to_wkt属性报错求助

我是geopandas新手，正参照教程在PySpark环境中处理自有shapefile。使用Jupyter（Python 3.7.6）执行代码时多次触发AttributeError: 'Polygon' object has no attribute 'to_wkt'错误，其他几何类型（如Point、Polygon Z）也出现相同错误。

执行代码 #

import os.path, json, io
import matplotlib.pyplot as plt
import matplotlib
matplotlib.style.use('ggplot')
matplotlib.rcParams['figure.figsize'] = (16, 20)

from retrying import retry # for exponential back down when calling TurboOverdrive API

import pyspark.sql.functions as func # resuse as func.coalace for example
from pyspark.sql.types import StringType, IntegerType, FloatType, DoubleType,DecimalType
from pyspark.sql import SparkSession

import pandas as pd
from geopandas import GeoDataFrame # Loading boundaries Data
from shapely.geometry import Point, Polygon, shape # creating geospatial data
from shapely import wkb, wkt # creating and parsing geospatial data
from ast import literal_eval as make_tuple # used to decode data from java

# Create SparkSession and attach Sparkcontext to it
spark = SparkSession.builder.appName("pyspark-geopandas").getOrCreate()
sc = spark.sparkContext

# Load the boundaries data
geo_df = GeoDataFrame.from_file('it_1km.shp')
geo_df.head()

geo_df.columns

geo_df.plot(column='PUNTI', categorical=True, legend=True)
plt.show()

wkts = map(lambda g: g.to_wkt() , geo_df.geometry)
#wkts[0]
type(geo_df.geometry)


geo_df.crs=('epsg:4326')
geo_df.crs

geo_df.geometry.area

wkts = map(lambda g: g.to_wkt() , geo_df.geometry)
type(geo_df.geometry)

geo_df.crs=('epsg:4326')
geo_df.crs

geo_df.geometry.area

geo_df['wkt'] = pd.Series(
        map(lambda geom: str(geom.to_wkt()), geo_df['geometry']),
        index=geo_df.index, dtype='string')

报错信息 #

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-14-f03fdc1f586f> in <module>
      1 geo_df['wkt'] = pd.Series(
      2         map(lambda geom: str(geom.to_wkt()), geo_df['geometry']),
----> 3         index=geo_df.index, dtype='string')

/opt/conda/lib/python3.7/site-packages/pandas/core/series.py in __init__(self, data, index, dtype, name, copy, fastpath)
    277                 data = data.to_dense()
    278             else:
--> 279                 data = com.maybe_iterable_to_list(data)
    280 
    281             if index is None:

/opt/conda/lib/python3.7/site-packages/pandas/core/common.py in maybe_iterable_to_list(obj)
    278     """
    279     if isinstance(obj, abc.Iterable) and not isinstance(obj, abc.Sized):
--> 280         return list(obj)
    281     return obj
    282 

<ipython-input-14-f03fdc1f586f> in <lambda>(geom)
      1 geo_df['wkt'] = pd.Series(
----> 2         map(lambda geom: str(geom.to_wkt()), geo_df['geometry']),
      3         index=geo_df.index, dtype='string')

AttributeError: 'Polygon' object has no attribute 'to_wkt'

这个错误的核心原因是你使用的Shapely版本过低——to_wkt()方法是在Shapely 1.8.0版本才新增的，而你的环境里的Shapely版本低于这个阈值，所以Shapely的几何对象（Polygon、Point等）没有这个方法。

有两种可行的解决办法：

方法1：升级Shapely到1.8.0及以上版本 #

在终端执行以下命令升级：

pip install --upgrade shapely

升级完成后重启Jupyter内核，再运行代码即可正常调用geom.to_wkt()方法。

方法2：使用兼容旧版本的wkt.dumps()替代 #

如果因为环境限制无法升级Shapely，可以改用你已经导入的shapely.wkt模块里的dumps()函数来转换几何对象为WKT字符串，修改代码如下：

# 替换原来的代码行
geo_df['wkt'] = pd.Series(
        map(lambda geom: wkt.dumps(geom), geo_df['geometry']),
        index=geo_df.index, dtype='string')

或者更简洁的Geopandas内置方式：

geo_df['wkt'] = geo_df.geometry.apply(wkt.dumps)

这种方式不需要修改Shapely版本，兼容所有支持Geopandas的旧版Shapely。

内容的提问来源于stack exchange，提问作者Luigi