Pandas-從具有嵌套串列串列的現有列創建動態列時出錯-有解無憂

我想從包含嵌套串列串列作為值的現有列創建兩列。

由 3 家公司參與者及其角色組成的記錄行：

**row 1** [{'roles': [{'type': 'director'}, {'type': 'founder'}, {'type': 'owner'}, {'type': 'real_owner'}], 'life': {'name': 'Lichun Du'}}]

**row 2** [{'roles': [{'type': 'board'}], 'life': {'name': 'Erik M?lgaard'}}, {'roles': [{'type': 'director'}, {'type': 'board'}, {'type': 'real_owner'}], 'life': {'name': 'Mikael Bodholdt Linde'}}, {'roles': [{'type': 'board'}, {'type': 'real_owner'}], 'life': {'name': 'Dorte B?cker Linde'}}]

**row 3** [{'roles': [{'type': 'director'}, {'type': 'real_owner'}], 'life': {'name': 'Kristian L?th Hougaard'}}, {'roles': [{'type': 'owner'}], 'life': {'name': 'WORLD JET HOLDING ApS'}}]

到目前為止，我已經嘗試過：

    responses['Role of Participant(s)'] = [element[0]['roles'] for element in responses['participants']]
    responses['Role of Participant(s)'] = responses['Role of Participant(s)'].apply(lambda x: ', '.join(t['type'] for t in x))
    responses['Name of Participant(s)'] = [element[0]['life']['name'] for element in responses['participants']]

這給了我以下輸出： Pandas-從具有嵌套串列串列的現有列創建動態列時出錯

它只回傳第一個參與者的角色和姓名

但是，我需要每行/記錄的所有參與者及其各自的角色，如下所示：

Pandas-從具有嵌套串列串列的現有列創建動態列時出錯

那么如何使用“***”作為每個行值的分隔符來實作這一點，就像上面的截圖一樣？ 請幫忙！！

更新： 這是資料框的 csv 版本：

participants
"[{'roles': [{'type': 'founder'}], 'life': {'name': 'Poul Erik Andersen'}}, {'roles': [{'type': 'director'}, {'type': 'board'}], 'life': {'name': 'Martin Ravn-Nielsen'}}, {'roles': [{'type': 'board'}], 'life': {'name': 'S?ren Haugaard'}}, {'roles': [{'type': 'board'}], 'life': {'name': 'Mads Dehlsen Winther'}}, {'roles': [{'type': 'founder'}], 'life': {'name': 'M  Ejendomme A/S'}}, {'roles': [{'type': 'founder'}], 'life': {'name': 'MILTON HOLDING HORSENS A/S'}}, {'roles': [{'type': 'accountant'}], 'life': {'name': 'EY Godkendt Revisionspartnerselskab'}}, {'roles': [{'type': 'owner'}], 'life': {'name': 'HUSCOMPAGNIET HOLDING A/S'}}]"
"[{'roles': [{'type': 'founder'}, {'type': 'director'}, {'type': 'board'}, {'type': 'real_owner'}], 'life': {'name': 'Rasmus Gert Hansen'}}, {'roles': [{'type': 'board'}], 'life': {'name': 'John Nyrup Larsen'}}, {'roles': [{'type': 'board'}], 'life': {'name': 'Ole Nidolf Larsen'}}, {'roles': [{'type': 'owner'}], 'life': {'name': 'RASMUS HANSEN HOLDING ApS'}}, {'roles': [{'type': 'accountant'}], 'life': {'name': 'DANSK REVISION SLAGELSE GODKENDT REVISIONSAKTIESELSKAB'}}]"
"[{'roles': [{'type': 'board'}], 'life': {'name': 'Berit Pedersen'}}, {'roles': [{'type': 'board'}], 'life': {'name': 'Sanne Kristine Sp?th'}}, {'roles': [{'type': 'real_owner'}], 'life': {'name': 'Kjeld Kirk Kristiansen'}}, {'roles': [{'type': 'director'}], 'life': {'name': 'Jesper Andersen'}}, {'roles': [{'type': 'board'}], 'life': {'name': 'Poul Hartvig Nielsen'}}, {'roles': [{'type': 'board'}], 'life': {'name': 'Nanna Birgitta Gudum'}}, {'roles': [{'type': 'board'}], 'life': {'name': 'Henrik Baag?e Fredel?kke'}}, {'roles': [{'type': 'board'}], 'life': {'name': 'Carsten Rasmussen'}}, {'roles': [{'type': 'board'}], 'life': {'name': 'Jesper Laursen'}}, {'roles': [{'type': 'board'}], 'life': {'name': 'John Hansen'}}, {'roles': [{'type': 'owner'}], 'life': {'name': 'LEGO A/S'}}, {'roles': [{'type': 'accountant'}], 'life': {'name': 'PRICEWATERHOUSECOOPERS STATSAUTORISERET REVISIONSPARTNERSELSKAB'}}]"

uj5u.com熱心網友回復：

你需要第二個for回圈而不是[0]

我使用普通函式而不是lambda使其更具可讀性。

首先是角色：

import pandas as pd

data = {'participants': 
[
    [{'roles': [{'type': 'director'}, {'type': 'founder'}, {'type': 'owner'}, {'type': 'real_owner'}], 'life': {'name': 'Lichun Du'}}],
    [{'roles': [{'type': 'board'}], 'life': {'name': 'Erik M?lgaard'}}, {'roles': [{'type': 'director'}, {'type': 'board'}, {'type': 'real_owner'}], 'life': {'name': 'Mikael Bodholdt Linde'}}, {'roles': [{'type': 'board'}, {'type': 'real_owner'}], 'life': {'name': 'Dorte B?cker Linde'}}],
    [{'roles': [{'type': 'director'}, {'type': 'real_owner'}], 'life': {'name': 'Kristian L?th Hougaard'}}, {'roles': [{'type': 'owner'}], 'life': {'name': 'WORLD JET HOLDING ApS'}}],
]
}

df = pd.DataFrame(data)

def get_roles(cell):
    
    results = []
    
    for item in cell:
        roles = []
        for role in item['roles']:
            roles.append(role['type'])
        results.append(",".join(roles))
    
    results = "***".join(results)

    return results

df['Role of Participant(s)'] = df['participants'].apply(get_roles)

print(df[['Role of Participant(s)']].to_string())

結果：

                                 Role of Participant(s)
0                     director,founder,owner,real_owner
1  board***director,board,real_owner***board,real_owner
2                           director,real_owner***owner

現在你可以嘗試寫成 lambda

df['Role of Participant(s)'] = df['participants'].apply(lambda cell:"***".join([",".join(role['type'] for role in item['roles']) for item in cell]))

名稱相似：

def get_names(cell):
    
    results = []
    
    for item in cell:
        results.append(item['life']['name'])
    
    results = "***".join(results)

    return results

df['Name of Participant(s)'] = df['participants'].apply(get_names)

現在有了 lambda

df['Name of Participant(s)'] = df['participants'].apply(lambda cell:"***".join(item['life']['name'] for item in cell))

編輯：

將兩列合二為一apply并跳過具有角色的參與者的版本director

import pandas as pd

data = {'participants': 
[
    [{'roles': [{'type': 'director'}, {'type': 'founder'}, {'type': 'owner'}, {'type': 'real_owner'}], 'life': {'name': 'Lichun Du'}}],
    [{'roles': [{'type': 'board'}], 'life': {'name': 'Erik M?lgaard'}}, {'roles': [{'type': 'director'}, {'type': 'board'}, {'type': 'real_owner'}], 'life': {'name': 'Mikael Bodholdt Linde'}}, {'roles': [{'type': 'board'}, {'type': 'real_owner'}], 'life': {'name': 'Dorte B?cker Linde'}}],
    [{'roles': [{'type': 'director'}, {'type': 'real_owner'}], 'life': {'name': 'Kristian L?th Hougaard'}}, {'roles': [{'type': 'owner'}], 'life': {'name': 'WORLD JET HOLDING ApS'}}],
]
}

df = pd.DataFrame(data)

def get_names_and_roles(cell):
    
    all_names = []
    all_roles = []
    
    for item in cell:
        name = item['life']['name']
        roles = [role['type'] for role in item['roles']]

        if 'director' not in roles:
            all_names.append(name)
            all_roles.append(",".join(roles))
    
    all_names = "***".join(all_names)
    all_roles = "***".join(all_roles)

    return pd.Series([all_names, all_roles])


df[ ['Name of Participant(s)', 'Role of Participant(s)'] ] = df['participants'].apply(get_names_and_roles)

print(df[ ['Name of Participant(s)', 'Role of Participant(s)'] ].to_string())

結果：

               Name of Participant(s)    Role of Participant(s)
0                                                              
1  Erik M?lgaard***Dorte B?cker Linde  board***board,real_owner
2               WORLD JET HOLDING ApS                     owner

轉載請註明出處，本文鏈接：https://www.uj5u.com/gongcheng/358224.html

標籤：Python 熊猫数据框 pandas-groupby 嵌套列表

上一篇：使PandasDataframe的最小n列為零的清潔方法

下一篇：Firebase實時資料庫正在Unity3d中添加額外的結果元素