我的数据框中有一个"列",在最佳条件下,如下所示:
Client: Stack Overflow Order Num: 123456 Account From: 3656645654 Account to: 546546578
我想将此列拆分为几列,例如:
'Client','Order Num', 'Account From','Account to'
但在某些情况下,我在列中没有客户、订单数和帐户
我是这样做的:
for x in len(df.columns):
if 'Client' in df.loc[x,'Columnn']:
df.loc[x,'Client'] = str(df.loc[x,'Column']).split('Client: ')[1]
if 'Order Num' in df.loc[x,'Client']:
df.loc[x,'Client'] = str(df.loc[x,'Client']).split('Order Num: ')[0]
if 'Account From' in df.loc[x,'Client']:
df.loc[x,'Client'] = str(df.loc[x,'Client']).split('Account From: ')[0]
if 'Account to' in df.loc[x,'Client']:
df.loc[x,'Client'] = str(df.loc[x,'Client']).split('Account to: ')[0]
else:
df.loc[x,'Client'] = ''
对于我要创建的所有列,依此类推。
这部分剧本差不多有40行,速度很慢。
你有更"熊猫"的解决方案吗?
使用字符串访问器尝试此操作,使用正则表达式.str
和命名组extract
:
df['col1'].str.extract('Client: (?P<Client>.*) Order Num: (?P<OrderNum>.*) Account From: (?P<AccountFrom>.*) Account to: (?P<AccountTo>.*)')
输出:
Client OrderNum AccountFrom AccountTo
0 Stack Overflow 123456 3656645654 546546578