使用函数在df中添加列

Question

      Date             Visitor  V_PTS                 Home  H_PTS  \
0 2012-10-30 19:00:00  Washington Wizards     84  Cleveland Cavaliers     94   
1 2012-10-30 19:30:00    Dallas Mavericks     99   Los Angeles Lakers     91   
2 2012-10-30 20:00:00      Boston Celtics    107           Miami Heat    120   
3 2012-10-31 19:00:00    Sacramento Kings     87        Chicago Bulls     93   
4 2012-10-31 19:30:00     Houston Rockets    105      Detroit Pistons     96   

试图添加到一个搜集到的数据集上，对NBA比赛的出勤率进行分析。我正在尝试添加一些列，比如arenaplayed和capacity。下面是我为添加竞技场而编写的函数的一部分。有没有更好的办法？我在datetime中有日期，所以如何正确地提取年份，以便将正确的竞技场分配给在过去几年中建造了较新竞技场的球队（萨克拉门托国王队）。另外，有没有增加球场容量，一石二鸟，而不是创造另一个功能？你知道吗

def label_arena (hometeam):
    if hometeam == 'Toronto Raptors' :
        return 'Air Canada Centre'
    if hometeam == 'Miami Heat' :
        return 'American Airlines Arena'
    if hometeam == 'Dallas Mavericks' :
        return 'American Airlines Center'
    if hometeam == 'Orlando Magic' :
        return 'Amway Center'
    if hometeam == 'San Antonio Spurs' :
        return 'AT&T Center'
    if hometeam == 'Indiana Pacers' :
        return 'Bankers Life Fieldhouse'
    if hometeam == 'Brooklyn Nets' :
        return 'Barclays Center'
    if hometeam == 'Milwaukee Bucks' :
        return 'Bradley Center'
    if hometeam == 'Washington Wizards' :
        return 'Capital One Arena'
    if hometeam == 'Oklahoma City Thunder' :
        return 'Chesapeake Energy Arena'
    if hometeam == 'Memphis Grizzlies' :
        return 'FedExForum'
    if hometeam == 'Sacramento Kings' and df['Date'] < 2016:
        return 'Sleep Train Arena'
    if hometeam == 'Sacramento Kings' and df['Date'] > 2016:
        return 'Golden 1 Center'

Answer 1

这就是简化逻辑的方法：

import pandas as pd

df = pd.DataFrame({'Date': ['2012-10-30', '2012-10-30', '2012-10-30',
                            '2012-10-31', '2017-10-31'],
                   'Home': ['Toronto Raptors', 'Los Angeles Lakers', 'Miami Heat',
                            'Sacramento Kings', 'Sacramento Kings']})

df['Date'] = pd.to_datetime(df['Date'])

d = {'Toronto Raptors': 'Air Canada Centre',
     'Los Angeles Lakers': 'Staples Center',
     'Miami Heat': 'American Airlines Arena'}

# general criteria
df['Arena'] = df['Home'].map(d)

# custom criteria
df.loc[(df['Home'] == 'Sacramento Kings') &
       (df['Date'].dt.year < 2016), 'Arena'] = 'Sleep Train Arena'
df.loc[(df['Home'] == 'Sacramento Kings') &
       (df['Date'].dt.year >= 2016), 'Arena'] = 'Golden 1 Center'

print(df)

        Date                Home                    Arena
0 2012-10-30     Toronto Raptors        Air Canada Centre
1 2012-10-30  Los Angeles Lakers           Staples Center
2 2012-10-30          Miami Heat  American Airlines Arena
3 2012-10-31    Sacramento Kings        Sleep Train Arena
4 2017-10-31    Sacramento Kings          Golden 1 Center

Answer 2

如果您不反对numpy，下面是一个使用^{}的方法：

import numpy as np

conditions = [
    df['Home'] == 'Toronto Raptors',
    df['Home'] == 'Miami Heat',
    df['Home'] == 'Dallas Mavericks',
    ...
    (df['Home'] == 'Sacramento Kings') & (df['Date'].dt.year < 2016),
    (df['Home'] == 'Sacramento Kings') & (df['Date'].dt.year > 2016)]

choices = [
        'Air Canada Centre',
        'American Airlines Arena',
        'American Airlines Center',
        ...
        'Sleep Train Arena',
        'Golden 1 Center']

df['arena'] = np.select(conditions, choices)

请注意，要使df['Date']条件起作用，需要将df['Date']设置为日期时间序列（如果尚未设置，可以通过df['Date'] = pd.to_datetime(df['Date'])执行此操作）

Answer 3

import pandas as pd

home_arenas_capacities = pd.DataFrame([
     ['Toronto Raptors', 'Air Canada Centre', 20511],
     ['Miami Heat', 'American Airlines Arena', 19600],
     ...
    ]) 

df.merge(home_arenas_capacities, on='Home')

对于萨克拉门托国王队，您希望合并“Home”和“Date”>；2016，这可能需要您创建一个临时列，然后df.merge(..., on=['Home','Date_GE_2016'])并删除“Date\u GE\u 2016”列。你知道吗

但更简单的方法是增加一列“季节”=“2015-16”，“2016-17”。随着您的数据库越来越大，您似乎需要它。（对于游戏数据库，您可以从'Date'值自动提取'Season'。对于“home\u arenas\u capacities”数据框，您需要手动编辑它）。你知道吗