拆分字符串的有效方法，其中单个单词由大写或数字分隔，并将其应用于整个数据帧列

Question

我有一个数据帧，其中团队名称以字符串形式给出。拆分字符串以获得单个团队名称的规则是

1. 每个大写的角色都会组建一个新团队
2. 除非前面的
3. 第一个非数字或非字母字符后没有团队名称

我的解决方案如下。它有效，但我正在寻找改进的指针

1. 整个拆分是否可以作为reg ex完成
2. 任何比使用更快的都适用。我看到了一些关于使用并行版本加速应用程序的注意事项，但不知道是否有更快的方法可以做到这一点

import pandas as pd
import numpy as np
import re
def splitDfCap(next_row):
new_team = ''
team_list = []
last_char_num = False
for next_char in next_row.Teams:
stop_current = re.match('[A-Z]', next_char) or (re.match('[0-9]', next_char) and not(last_char_num))
if stop_current and len(new_team) > 0:
team_list.append(new_team)
new_team = ''

if re.match('[a-zA-Z0-9]', next_char):
new_team = new_team + next_char
last_char_num = re.match('[0-9]', next_char)
else:
break

if len(new_team) > 0:
team_list.append(new_team)

return team_list

#Setup
df = pd.DataFrame({'Population' : [1400, 100, 1000], 
'Teams' : ['YankeesMetsBrooklyn[]', 'Rays', 'GiantsAthletics49ers'], 
'Ad Market' : [10400, 2000, 8400]},
index = ['New York Area', 'Tampa Bay', 'SF Bay Area'])
#Create list of teams and then explode
df['Team List'] = df.apply(splitDfCap, axis='columns')
df = df.rename_axis(index='Region')
.explode('Team List')
.rename(columns={'Team List': 'Team'})
.drop(columns=['Teams'])
print(df)

Output is
Population  Ad Market       Team
Region                                         
New York Area        1400      10400    Yankees
New York Area        1400      10400       Mets
New York Area        1400      10400   Brooklyn
Tampa Bay             100       2000       Rays
SF Bay Area          1000       8400     Giants
SF Bay Area          1000       8400  Athletics
SF Bay Area          1000       8400      49ers

Answer 1

import pandas as pd
df = pd.DataFrame({'Population' : [1400, 100, 1000], 
'Teams' : ['YankeesMetsBrooklyn[]', 'Rays', 'GiantsAthletics49ers'], 
'Ad Market' : [10400, 2000, 8400]},
index = ['New York Area', 'Tampa Bay', 'SF Bay Area'])
df['Teams']=df.Teams.str.findall('(?:[A-Z]|d+)[^A-Zd+W]*')
df = df.explode('Teams')
df.reset_index(inplace=True)
df.columns = ['Region','Population','Ad Market','Team']