如何将 DataFrame 中的每一行/单元格值转换为 pandas 中的字典列表？

Posted 2023-03-11

【中文标题】如何将 DataFrame 中的每一行/单元格值转换为 pandas 中的字典列表？

【英文标题】：How to convert each row/cell values from a DataFrame to a list of dictionaries in pandas?

【发布时间】：2022-01-19 02:04:14

【问题描述】：

我在下面有一个熊猫数据框：

df_input = pd.DataFrame(
         'Domain':['www.google.com','www.apple.com','www.amazon.com'],
         'Description':['The company’s product portfolio includes Googl...','Apple is a multinational corporation that desi...','Amazon is an international e-commerce website ...'],
         'About Us':['Google is a multinational corporation that spe...','Apple is a multinational corporation that desi...','Amazon is an e-commerce website for consumers,...'],
         'Founded':[1998, 1976, 1994],
         'Country':['United States','United States','United States'])

我想按如下方式传输：

["properties": ["name": "Description","value": "The company’s product portfolio includes Google Search, which provides users with access to information online; Knowledge Graph that allows users to search for things, people, or places as well as builds systems recognizing speech and understanding",
            "name": "Domain","value": "www.google.com",
            "name": "About Us", "value": "Google is a multinational corporation that specializes in Internet-related services and products.",
            "name": "Founded", "value": 1998,
            "name": "Country", "value":"United States"],

"properties": ["name": "Description","value": "Apple is a multinational corporation that designs, manufactures, and markets mobile communication and media devices, personal computers, portable digital music players, and sells a variety of related software, services, peripherals, networking solutions, and third-party digital content and applications.",
            "name": "Domain","value": "www.apple.com",
            "name": "About Us", "value": "Apple is a multinational corporation that designs, manufactures, and markets consumer electronics, personal computers, and software.",
            "name": "Founded", "value": 1976,
            "name": "Country", "value":"United States"],

"properties": ["name": "Description","value": "Amazon is an international e-commerce website for consumers, sellers, and content creators. It offers users merchandise and content purchased for resale from vendors and those offered by third-party sellers.",
            "name": "Domain","value": "www.amazon.com",
            "name": "About Us", "value": "Amazon is an e-commerce website for consumers, sellers, and content creators.",
            "name": "Founded", "value": 1994,
            "name": "Country", "value":"United States"]]

如何编写循环来执行此操作？

【参考方案1】：

您只需遍历行。

import pandas as pd

df_input = pd.DataFrame(
         'Domain':['www.google.com','www.apple.com','www.amazon.com'],
         'Description':['The companys product portfolio includes Googl...','Apple is a multinational corporation that desi...','Amazon is an international e-commerce website ...'],
         'About Us':['Google is a multinational corporation that spe...','Apple is a multinational corporation that desi...','Amazon is an e-commerce website for consumers,...'],
         'Founded':[1998, 1976, 1994],
         'Country':['United States','United States','United States'])

data = []
for row in df_input.iterrows():
    props = []
    for key,val in zip(df_input.columns, row[1].values ):
        props.append( 'name':key, 'value':val )
    data.append( 'properties': props )
from pprint import pprint
pprint(data)

输出：

['properties': ['name': 'Domain', 'value': 'www.google.com',
                 'name': 'Description',
                  'value': 'The companys product portfolio includes Googl...',
                 'name': 'About Us',
                  'value': 'Google is a multinational corporation that spe...',
                 'name': 'Founded', 'value': 1998,
                 'name': 'Country', 'value': 'United States'],
 'properties': ['name': 'Domain', 'value': 'www.apple.com',
                 'name': 'Description',
                  'value': 'Apple is a multinational corporation that desi...',
                 'name': 'About Us',
                  'value': 'Apple is a multinational corporation that desi...',
                 'name': 'Founded', 'value': 1976,
                 'name': 'Country', 'value': 'United States'],
 'properties': ['name': 'Domain', 'value': 'www.amazon.com',
                 'name': 'Description',
                  'value': 'Amazon is an international e-commerce website ...',
                 'name': 'About Us',
                  'value': 'Amazon is an e-commerce website for consumers,...',
                 'name': 'Founded', 'value': 1994,
                 'name': 'Country', 'value': 'United States']]

【讨论】：

谢谢蒂姆！这正是我所希望的。很高兴知道 iterrows() 可以解决问题。

我确实需要给你一个强制性警告，即遍历 pandas 中的行或列很慢。如果可能，最好找到批量操作。在这种情况下，我认为没有其他选择。