使用Pandas读取CSV时，如何在迭代中按列名访问数据？

问题场景 #

读取CSV文件后，希望在循环中通过列名（如product_url）访问每行数据，尝试了以下代码但无法实现预期效果：

import pandas as pd

products = pd.read_csv(f'./storage/temp/{file_name}', delimiter=',', on_bad_lines='skip').to_dict()
        
for index, product in products:
    print(product.product_url)

CSV示例数据：

product_url,image_url,categories,attributes,description,product_name,reviews,stock,sku,product_gift,price,price_with_discount,categories_urls
http://localhost:8000/?product=aw-bellies,http://localhost:8000/wp-content/uploads/2022/10/red-as-454-aw-11-original-imaeebfwsdf6jdf6-324x324.jpeg,Footwear,"""[{""""Ideal For"""":""""Women""""}","{""""Occasion"""":""""Casual""""}","{""""Color"""":""""Red""""}]""""","""Key Features of AW Bellies Sandals Wedges Heel Casuals",AW Bellies Price: Rs. 499 Material: Synthetic Lifestyle: Casual Heel Type: Wedge Warranty Type: Manufacturer Product Warranty against manufacturing defects: 30 days Care instructions: Allow your pair of shoes to air and de-odorize at regular basis; use shoe bags to prevent any stains or mildew; dust any dry dirt from the surface using a clean cloth; do not use polish or shiner,"Specifications of AW Bellies General Ideal For Women Occasion Casual Shoe Details Color Red Outer Material Patent Leather Heel Height 1 inch Number of Contents in Sales Package Pack of 1 In the Box One Pair Of Shoes""""","""AW Bellies""""",,1,2,,499,499,http://localhost:8000/?product_cat=footwear

期望的访问方式：

for product in products:
   print(product.product_url)

错误原因 #

1. to_dict()默认参数为'dict'，会将DataFrame转换为以列名为键、对应列数据为值的字典，并非按行存储的结构，因此循环时无法拿到每行数据。
2. 即使转换为行字典，普通字典也不能用.语法访问键，需使用[]语法。

解决方案 #

提供三种可行实现方式，按需选择：

方法1：使用itertuples()（支持.语法访问列名） #

itertuples()返回每行数据的命名元组，可直接通过列名作为属性访问：

import pandas as pd

df = pd.read_csv(f'./storage/temp/{file_name}', delimiter=',', on_bad_lines='skip')
for product in df.itertuples(index=False):
    print(product.product_url)

• index=False表示不返回行索引，仅返回行数据。

方法2：转换为行字典列表（用[]访问键） #

用to_dict('records')将DataFrame转换为每行对应一个字典的列表，循环时通过键名访问：

import pandas as pd

products = pd.read_csv(f'./storage/temp/{file_name}', delimiter=',', on_bad_lines='skip').to_dict('records')
for product in products:
    print(product['product_url'])

方法3：使用iterrows()（返回索引和行Series） #

iterrows()返回每行的索引和Series对象，Series可用.或[]访问列名：

import pandas as pd

df = pd.read_csv(f'./storage/temp/{file_name}', delimiter=',', on_bad_lines='skip')
for index, product in df.iterrows():
    print(product.product_url)  # 或 product['product_url']

内容的提问来源于stack exchange，提问作者DaniD