Python Pandas：将通过API调用检索的非类型数组从列表转换为字符串（以启用透视表的使用）

我查询一个内部数据库，检索数据，并创建一个看起来类似于以下内容的dataframe

df

：

import pandas as pd
df = pd.DataFrame({'issue_key':['MED-187', 'MED-188', 'MED-190', 'MED-191'],
                    'creator': ['Smith, J', 'Williams, S', 'Wilson, T', 'Smith, J'],
                    'manufacturer': ['Mercedes', 'Audi', 'Mercedes', 'BMW'],
                    'department': [['Sales'], ['Finance'], ['Operations'], ['Sales']],
                    'expense': [29181, 12809, 837, 2817]})

issue_key    creator        manufacturer      department        expense
MED-187      Smith, J       Mercedes          [Sales]           29181
MED-188      Williams, S    Audi              [Finance]         12809
MED-190      Wilson, T      Mercedes          [Operations]      837
MED-191      Smith, J       BMW               [Sales]           2817

import numpy as np
pivoted_data = pd.pivot_table(data=df, index='department', values='expense', aggfunc=np.mean)

df['new_department'] = [','.join(map(str, l)) for l in df['department']]

`df['activity'].unique()`

然后，我尝试使用以下方法来透视数据：

import pandas as pd
df = pd.DataFrame({'issue_key':['MED-187', 'MED-188', 'MED-190', 'MED-191'],
                    'creator': ['Smith, J', 'Williams, S', 'Wilson, T', 'Smith, J'],
                    'manufacturer': ['Mercedes', 'Audi', 'Mercedes', 'BMW'],
                    'department': [['Sales'], ['Finance'], ['Operations'], ['Sales']],
                    'expense': [29181, 12809, 837, 2817]})

issue_key    creator        manufacturer      department        expense
MED-187      Smith, J       Mercedes          [Sales]           29181
MED-188      Williams, S    Audi              [Finance]         12809
MED-190      Wilson, T      Mercedes          [Operations]      837
MED-191      Smith, J       BMW               [Sales]           2817

import numpy as np
pivoted_data = pd.pivot_table(data=df, index='department', values='expense', aggfunc=np.mean)

df['new_department'] = [','.join(map(str, l)) for l in df['department']]

`df['activity'].unique()`

但是，我得到了这个错误：

TypeError: unhashable type: 'list'

因此，我使用以下命令将department列从我认为是一个列表（稍后将详细介绍）转换为一个字符串：

import pandas as pd
df = pd.DataFrame({'issue_key':['MED-187', 'MED-188', 'MED-190', 'MED-191'],
                    'creator': ['Smith, J', 'Williams, S', 'Wilson, T', 'Smith, J'],
                    'manufacturer': ['Mercedes', 'Audi', 'Mercedes', 'BMW'],
                    'department': [['Sales'], ['Finance'], ['Operations'], ['Sales']],
                    'expense': [29181, 12809, 837, 2817]})

issue_key    creator        manufacturer      department        expense
MED-187      Smith, J       Mercedes          [Sales]           29181
MED-188      Williams, S    Audi              [Finance]         12809
MED-190      Wilson, T      Mercedes          [Operations]      837
MED-191      Smith, J       BMW               [Sales]           2817

import numpy as np
pivoted_data = pd.pivot_table(data=df, index='department', values='expense', aggfunc=np.mean)

df['new_department'] = [','.join(map(str, l)) for l in df['department']]

`df['activity'].unique()`

正如预期的那样，其结果是：

df = pd.DataFrame({'issue_key':['MED-187', 'MED-188', 'MED-190', 'MED-191'],
                    'creator': ['Smith, J', 'Williams, S', 'Wilson, T', 'Smith, J'],
                    'manufacturer': ['Mercedes', 'Audi', 'Mercedes', 'BMW'],
                    'department': [['Sales'], ['Finance'], ['Operations'], ['Sales']],
                    'expense': [29181, 12809, 837, 2817],
                    'new_department': ['Sales', 'Finance', 'Operations', 'Sales']
                  })

issue_key    creator        manufacturer      department        expense     new_department
MED-187      Smith, J       Mercedes          [Sales]           29181       Sales
MED-188      Williams, S    Audi              [Finance]         12809       Finance
MED-190      Wilson, T      Mercedes          [Operations]      837         Operations
MED-191      Smith, J       BMW               [Sales]           2817        Sales

  

array(["[<JIRA CustomFieldOption:  value='Sales', id='174'>]",
       "[<JIRA CustomFieldOption:  value='Finance', id='179'>]",
       "[<JIRA CustomFieldOption:  value='Operations', id='102'>]",
       'None'], dtype=object)

现在，我能够成功地透视数据

但是在生产数据上，我得到以下错误：

df['new_department'] = [','.join(map(str, l)) for l in df['department']]

TypeError: 'NoneType' object is not iterable

在研究如何从API返回数据时，我调用了以下命令：

import pandas as pd
df = pd.DataFrame({'issue_key':['MED-187', 'MED-188', 'MED-190', 'MED-191'],
                    'creator': ['Smith, J', 'Williams, S', 'Wilson, T', 'Smith, J'],
                    'manufacturer': ['Mercedes', 'Audi', 'Mercedes', 'BMW'],
                    'department': [['Sales'], ['Finance'], ['Operations'], ['Sales']],
                    'expense': [29181, 12809, 837, 2817]})

issue_key    creator        manufacturer      department        expense
MED-187      Smith, J       Mercedes          [Sales]           29181
MED-188      Williams, S    Audi              [Finance]         12809
MED-190      Wilson, T      Mercedes          [Operations]      837
MED-191      Smith, J       BMW               [Sales]           2817

import numpy as np
pivoted_data = pd.pivot_table(data=df, index='department', values='expense', aggfunc=np.mean)

df['new_department'] = [','.join(map(str, l)) for l in df['department']]

`df['activity'].unique()`

其结果是：

df = pd.DataFrame({'issue_key':['MED-187', 'MED-188', 'MED-190', 'MED-191'],
                    'creator': ['Smith, J', 'Williams, S', 'Wilson, T', 'Smith, J'],
                    'manufacturer': ['Mercedes', 'Audi', 'Mercedes', 'BMW'],
                    'department': [['Sales'], ['Finance'], ['Operations'], ['Sales']],
                    'expense': [29181, 12809, 837, 2817],
                    'new_department': ['Sales', 'Finance', 'Operations', 'Sales']
                  })

issue_key    creator        manufacturer      department        expense     new_department
MED-187      Smith, J       Mercedes          [Sales]           29181       Sales
MED-188      Williams, S    Audi              [Finance]         12809       Finance
MED-190      Wilson, T      Mercedes          [Operations]      837         Operations
MED-191      Smith, J       BMW               [Sales]           2817        Sales

  

array(["[<JIRA CustomFieldOption:  value='Sales', id='174'>]",
       "[<JIRA CustomFieldOption:  value='Finance', id='179'>]",
       "[<JIRA CustomFieldOption:  value='Operations', id='102'>]",
       'None'], dtype=object)

数组（[“[]”，
"[]",
"[]",
“无”]，数据类型=对象）

API返回的结果显示

None

，因此错误消息完全有意义

你知道如何克服这个问题吗？是否可以以某种方式转换

None

类型，以便遍历列表？或者，直接访问数组中的值

我已经调查了几个小时，无法解决这个问题

---

提前感谢您的帮助

如果问题是“部门”列中的rogue

None

值，则可以使用列表中的条件语句来处理它：

df['new_department'] = [','.join(map(str, l)) if l is not None else 'NA' for l in df['department']]

或者更一般地说：

unwanted_values = [None,np.NaN]
df['new_department'] = ['NA' if l in unwanted_values else ','.join(map(str, l)) for l in df['department']]