在Dataframe中访问字典

Question

我有以下数据帧(有1324行)：enter image description here

我需要找出哪些城市适合外卖('Take_ out '：属性字典中的True )

Answer 1

为了得到这个答案，我首先创建了一个用于测试的虚拟DataFrame：

import numpy as np
import pandas as pd
# create a dictionary list
d = list({'Take-out': True} for x in np.arange(10))
ddf = pd.Series(d, name='attributes')
ddf = pd.DataFrame(ddf)
ddf.index.name = 'cities'
print(ddf)

这将生成一个与您的图像中类似的DataFrame。

接下来，迭代DataFrame，访问“attributes”列，如下所示：

# cities buffer will hold successes
cities = []
# iterate over the list of dictionaries:
for i, each in enumerate(ddf['attributes']):
    # check if the keys is in that dictionary, if so, keep the city name
    if 'Take-out' in ddf['attributes'][i].keys():
        # the index is named 'cities' and each position is a city name, so:
        cities.append(ddf.index[i])
print(cities)

Answer 2

字典并不是你真正想要存储在DataFrame中的东西，但是你可以试试：

df = df.assign(**df.attributes.dropna().apply(pd.Series))
df[df["Take_out"] == True]

这与这里描述的思想大致相同：Unpack dictionary from Pandas Column。解开字典后，我们可以像往常一样用"Take_out“== True选择行。