python - 获取存储在数据框列中的键值？

Question

我有一个如下所示的数据框：

user_id event_desc
100013  {'firebase_screen': '/product-popup', 'banner_id': '9260'}
100022  {'firebase_screen': '/product-popup', 'banner_id': '9152'}
100076  {'firebase_screen_id': '-1531185473150377061'}
100076  {'banner_id': '8567', 'firebase_screen': '/product-popup'}

我正在尝试获取键banner_id 的值。

所以，输出如下：

user_id event_desc                                            banner_id
100013  {'firebase_screen': '/product-popup', 'banner_id': '9260'}  9260
100022  {'firebase_screen': '/product-popup', 'banner_id': '9152'}  9151
100076  {'firebase_screen_id': '-1531185473150377061'}               Nan
100076  {'banner_id': '8567', 'firebase_screen': '/product-popup'}  8567

我正在使用下面的代码。

df['banner_id'] = df['event_desc'].apply(lambda x: x.get('banner_id'))

它给出了错误：

“str”对象没有属性“get”

标签： python

Answer 1

只要这些不存储为字符串，它就可以工作：

import pandas as pd

df = pd.DataFrame([
[100013,  {'firebase_screen': '/product-popup', 'banner_id': '9260'}],
[100022,  {'firebase_screen': '/product-popup', 'banner_id': '9152'}],
[100076 , {'firebase_screen_id': '-1531185473150377061'}],
[100076 , {'banner_id': '8567', 'firebase_screen': '/product-popup'}]], columns =['user_id', 'event_desc'])


df['banner_id'] = df['event_desc'].apply(lambda x: x.get('banner_id'))   

我的猜测是，根据您的错误，该列不是字典列，而是字符串列：

import pandas as pd

df = pd.DataFrame([
[100013,  "{'firebase_screen': '/product-popup', 'banner_id': '9260'}"],
[100022,  "{'firebase_screen': '/product-popup', 'banner_id': '9152'}"],
[100076 , "{'firebase_screen_id': '-1531185473150377061'}"],
[100076 , "{'banner_id': '8567', 'firebase_screen': '/product-popup'}"]], columns =['user_id', 'event_desc'])


df['banner_id'] = df['event_desc'].apply(lambda x: x.get('banner_id'))

AttributeError: 'str' object has no attribute 'get'

因此，如果是这种情况，您将需要使用：

import pandas as pd
import ast

df = pd.DataFrame([
[100013,  "{'firebase_screen': '/product-popup', 'banner_id': '9260'}"],
[100022,  "{'firebase_screen': '/product-popup', 'banner_id': '9152'}"],
[100076 , "{'firebase_screen_id': '-1531185473150377061'}"],
[100076 , "{'banner_id': '8567', 'firebase_screen': '/product-popup'}"]], columns =['user_id', 'event_desc'])


df['banner_id'] = df['event_desc'].apply(lambda x: ast.literal_eval(x).get('banner_id')) 

额外的：

import pandas as pd
import ast


def eval_code(code):
    parsed = ast.parse(code, mode='eval')
    fixed = ast.fix_missing_locations(parsed)
    compiled = compile(fixed, '<string>', 'eval')
    return eval(compiled)


df = pd.DataFrame([
[100013,  "{'firebase_screen': '/product-popup', 'banner_id': '9260'}"],
[100022,  "{'firebase_screen': '/product-popup', 'banner_id': '9152'}"],
[100076 , "{'firebase_screen_id': '-1531185473150377061'}"],
[100076 , "{'banner_id': '8567', 'firebase_screen': '/product-popup'}"]], columns =['user_id', 'event_desc'])


df['banner_id'] = df['event_desc'].apply(lambda x: eval_code(x).get('banner_id'))