Pandas批量更新帐户字符串

我的组织拥有由多个字段组合而成的帐号。最后一个字段始终为4个字符(通常为0000)

Org           Account
 01       01-123-0000
 01       01-456-0000
 02       02-789-0000
 02       02-456-0000
 03       03-987-0000
 03       03-123-1234

我还有一个字典映射,说明最后一个组件应该有多少个字符。

MAP = {'01': 4, '02': 3, '03': 3}

然而,组织03也有特殊的映射:

D03_SPECIAL_MAP = {'0000': '012', '1234': '123'}

我更新最后一个组件的代码是:

for i,r in df.iterrows():
    updated = False # Keep track if we have updated this row

    # Split off last component from the rest of the account
    Acct, last_comp = r['Account'].rsplit('-',1)

    # Check if we need to update code length and the code length does not match
    if r['Org'] in MAP and len(last_comp) != MGMT_MAP[r['Org']]:
        df.at[i,'Account'] = '-'.join(Acct) + "-" + last_comp.zfill(MAP[r['Org']])
        updated = True

    # Special mapping for Org 03
    if r['Org'] =='03' and last_comp in D03_SPECIAL_MAP.keys():
        df.at[i,'Account'] = '-'.join(Acct) + "-" + D03_SPECIAL_MAP[last_comp]
        updated = True

    if not updated: # Join Default if we have not hit either of he conditions above
        df.at[i,'Account'] = '-'.join(Acct) + "-" + last_comp

其输出将是:

Org           Account
 01       01-123-0000
 01       01-456-0000
 02       02-789-000
 02       02-456-000
 03       03-987-012
 03       03-123-123

我的代码按预期工作,只是这个过程检查每条记录有点慢。是否有方法在不使用的情况下执行相同的操作 df.iterrows() ?