如何在单独的数据帧列中从字典列表中的特定值创建列表?

Question

我有一个名为'new_api_df'的数据帧，其中有一个名为new_api_df['Categories']的列，其中包含字典列表：

[{'CallId': 22143866, 'BucketId': 1953, 'SectionId': 1256, 'BucketFullName': 'Categories.Filters.No Sale Made', 'Weight': 1.0}
, {'CallId': 22143866, 'BucketId': 2016, 'SectionId': 1255, 'BucketFullName': 'Categories.Imported.Objections', 'Weight': 3.0}
, {'CallId': 22143866, 'BucketId': 2017, 'SectionId': 1255, 'BucketFullName': 'Categories.Imported.Touting Benefits', 'Weight': 1.0}
]

我想获取每个"BucketFullName"并将这些值放入单独列的列表中，new_api_df['category_list']如下所示：

['Categories.Filters.No Sale Made', 'Categories.Imported.Objections', Categories.Imported.Touting Benefits']

我尝试使用列表理解，如下所示：

new_api_df['category_list'] =[item['BucketFullName'] for dictionary in new_api_df['Categories'] for item in dictionary]

但得到错误：ValueError: Length of values does not match length of index

我还尝试了应用和列表理解：new_api_df['category_list'] = new_api_df['Categories'].apply([item['BucketFullName'] for dictionary in new_api_df['Categories'] for item in dictionary])

但我收到以下错误：AttributeError: 'Categories.Filters.No Sale Made' is not a valid function for 'Series' object

我也试过：new_api_df['category_list'] = df['Categories'].apply(lambda x: x['BucketFullName'])

但得到错误：TypeError: list indices must be integers or slices, not str

new_api_df切片：

new_api_df.loc[0]：

Contact            {'Id': 22143866, 'Type': 'Call', 'WavPath': '...
RecordInfo         {'Id': 22143866, 'RowNumber': 1, 'TotalRowCoun...
Measures           {'ID': 22143866, 'TotalHoldCount': 0, 'Agitati...
Others             {'ConfidenceAverage': 69, 'SequenceID': None, ...
Sections           [{'CallId': 22143866, 'SectionId': 1041, 'Sect...
Categories         [{'CallId': 22143866, 'BucketId': 1953, 'Secti...
Scores             [{'CallId': 22143866, 'ScoreId': 399, 'ScoreNa...
ScoreComponents    [{'CallId': 22143866, 'ScoreComponentId': 4497...```

Answer 1

我想你想要

df=pd.DataFrame({'Categories':[{'CallId': 22143866, 'BucketId': 1953, 'SectionId': 1256, 'BucketFullName': 'Categories.Filters.No Sale Made', 'Weight': 1.0}
, {'CallId': 22143866, 'BucketId': 2016, 'SectionId': 1255, 'BucketFullName': 'Categories.Imported.Objections', 'Weight': 3.0}
, {'CallId': 22143866, 'BucketId': 2017, 'SectionId': 1255, 'BucketFullName': 'Categories.Imported.Touting Benefits', 'Weight': 1.0}
]})
df['category_list']=df['Categories'].apply(lambda x: x[0]['BucketFullName'])
print(df)
#                                          Categories  
#0  {'CallId': 22143866, 'BucketId': 1953, 'Sectio...   
#1  {'CallId': 22143866, 'BucketId': 2016, 'Sectio...   
#2  {'CallId': 22143866, 'BucketId': 2017, 'Sectio...   
#
#                          category_list  
#0       Categories.Filters.No Sale Made  
#1        Categories.Imported.Objections  
#2  Categories.Imported.Touting Benefits  

更新

请注意，这将为每个单元格创建一个列表。

df['category_list']=df['Categories'].apply(lambda x: [m_dict['BucketFullName'] for m_dict in x])

然后你可以使用DataFrame.explode

df = df.explode('category_list')
#(df['Categories'].apply(lambda x: [m_dict['BucketFullName'] for m_dict in x])                                                             
#                  .explode()) #check the explode serie