在python中，如何将结果保存到csv，AttributeError:“dict”对象没有“to_csv”属性

Question

如何将结果按顺序保存到csv？意思是添加额外的列解释什么是“添加”和什么是“删除”以及什么是“更改”

我尝试了diff.to_csv（'diff.csv'）并得到了这个错误

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: 'dict' object has no attribute 'to_csv'

这是python代码

from csv_diff import load_csv, compare
diff = compare(
    load_csv(open("list1.csv"), key="ean"),
    load_csv(open("list2.csv"), key="ean")
)
diff.to_csv('diff.csv')

额外信息

list1.csv

price, oldprice,title,brand,category,unit,ean,,
17,,VR BOX Virtual Reality 3D Glasses Bluetooth Game Remote Control For Phone Iphone,other,3d glasses,2023700513,272434,,
18,,3d Glasses,other,3d glasses,1493500513,2272434,,
22,,Zefas Active 3D GlassesFor SmartPhones,zefas,3d glasses,1342700513,2272456,,
22.98,,3D Glasses Circular Polarized Lenses for Polarized TV| 3D Cinemas,other,3d glasses,1992100513,2272483,,
25,,max3 pro,other,3d glasses,1904600513,2272432,,
28.27,,Red Blue Clip on Anaglyph Glasses,other,3d glasses,2068900513,2272453,,

list2.csv

price,oldprice,title,brand,category,unit,ean,,
22.98,,3D Glasses Circular Polarized Lenses for Polarized TV| 3D Cinemas,other,3d  glasses,1992100513,2272483,,
25,,max3 pro,other,3d glasses,1904600513,2272432,,
59,,Red-blue Cyan Anaglyph 3D,terratec,3d glasses,2103700513,2272428,,
65,,Sinogoodies Passive 3D GlassesFor SmartPhones,sinogoodies,3d glasses,1603700513,2272464,,
69.91,,G15-DLP 3D Active Shutter Glasses for DLP-LINK DLP LINK 3D for Optoma Sharp LG Acer BenQ Projectors,other,3d glasses,2039200513,227243,,
70,,Sinogoodies Passive 3D GlassesFor Multi,sinogoodies,3d glasses,1603600513,2272464,,

运行的屏幕结果

compare(
        load_csv(open("list1.csv"), key="ean"),
        load_csv(open("list2.csv"), key="ean")
    )

是

{'added': [{'price': '59', 'oldprice': '', 'title': 'Red-blue Cyan Anaglyph 3D',
 'brand': 'terratec', 'category': '3d glasses', 'unit': '2103700513', 'ean': '22
72428', '': ''}, {'price': '70', 'oldprice': '', 'title': 'Sinogoodies Passive 3
D GlassesFor Multi', 'brand': 'sinogoodies', 'category': '3d glasses', 'unit': '
1603600513', 'ean': '2272464', '': ''}, {'price': '69.91', 'oldprice': '', 'titl
e': 'G15-DLP 3D Active Shutter Glasses for DLP-LINK DLP LINK 3D for Optoma Sharp
 LG Acer BenQ Projectors', 'brand': 'other', 'category': '3d glasses', 'unit': '
2039200513', 'ean': '227243', '': ''}], 'removed': [{'price': '17', 'oldprice':
'', 'title': 'VR BOX Virtual Reality 3D Glasses Bluetooth Game Remote Control Fo
r Phone Iphone', 'brand': 'other', 'category': '3d glasses', 'unit': '2023700513
', 'ean': '272434', '': ''}, {'price': '18', 'oldprice': '', 'title': '3d Glasse
s', 'brand': 'other', 'category': '3d glasses', 'unit': '1493500513', 'ean': '22
72434', '': ''}, {'price': '22', 'oldprice': '', 'title': 'Zefas Active 3D Glass
esFor SmartPhones', 'brand': 'zefas', 'category': '3d glasses', 'unit': '1342700
513', 'ean': '2272456', '': ''}, {'price': '28.27', 'oldprice': '', 'title': 'Re
d Blue Clip on Anaglyph Glasses', 'brand': 'other', 'category': '3d glasses', 'u
nit': '2068900513', 'ean': '2272453', '': ''}], 'changed': [], 'columns_added':
[], 'columns_removed': []}

此解决方案不起作用

import pandas as pd
dataframe = pd.DataFrame(diff)
dataframe.to_csv("data.csv", header=True)

Answer 1

错误'dict' object has no attribute 'to_csv'表示diff变量是字典，因此没有名为.to_csv的方法

您可以尝试将diff强制转换为csv文件，但无法正常工作。您的diff当前是一个字典，其值是字典数组。此格式不适用于csv。如果使用Pandas包，则需要字典中的值是字符串的数组。下面是带有示例的docs

我建议您先考虑如何格式化diff，然后再转换为csv

一个可行的方法是从diff字典创建两个CSV。添加和删除的值实际上是csv的正确格式。以下是一个例子：

import csv

keys = diff['added'][0].keys()
with open('file_name.csv', 'w') as output_file:
  dict_writer = csv.DictWriter(output_file, keys)
  dict_writer.writeheader()
  dict_writer.writerows(diff['added'])

这些是csv包的docs

编辑：键应从添加的第一个元素中获取键

Answer 2

其他解决方案不起作用，因为您的数据没有正确的DataFrame结构。您需要首先从顶级键访问dict列表

import pandas

df1  = pandas.DataFrame(diff['added'])
df1['change'] = 'added'
df2 = pandas.DataFrame(diff['removed'])
df2['change'] = 'removed'

df = df1.append(df2)
df.to_csv('diff.csv')

更改每个评论请求

df1 = pandas.read_csv('list1.csv')
df1['version'] = 'list1'
df2 = pandas.read_csv('list2.csv')
df2['version'] = 'list2'

# keep only columns 'version', 'ean', 'price'
diff = df1.append(df2)[['version', 'ean', 'price']]
# keep only duplicated eans, which will only occur
# for eans in both original lists
diff = diff[diff['ean'].duplicated(keep=False)]
# perform a pivot https://pandas.pydata.org/pandas-docs/stable/user_guide/reshaping.html
diff = diff.pivot(index='ean', columns='version', values='price')

# back to a normal dataframe
diff = diff.reset_index()
diff.columns.name = None

# rename columns and keep only what we want
diff = diff.rename(columns={'list1': 'price1', 'list2': 'price2'})[['ean', 'price1', 'price2']]
diff['difference'] = diff['price2'] - diff['price1']

Answer 3

您可以使用由pandas提供的to_csv方法：

import pandas as pd 

df = pd.DataFrame(dict([ (k,pd.Series(v)) for k,v in diff.items() ]))
df.to_csv('diff.csv')