为pandas中的不同行分配掩码值

Question

我有一个pandas数据框架

df =     code       mapper       version  country  range    month              value
1       ABC321       ABC          Alpha     USA     High   2021-10              8.0
6       ABC321       ABC          Alpha     USA     High   2021-11              1.0
2       ABC321       PQS          Beta      IND     LOW    2021-10              0.0
3       ABC321       TRR          Delta     MEX     LOW    2021-10              1.0
4       ABC321       TRR          Delta     MEX     LOW    2021-11              3.0

我基于条件屏蔽行并使值为零

mask = (
(df.mapper == 'ABC')
& (df.version == 'Alpha')
& (df.country == 'USA')
& (df.range == 'High')
& (df.month == '2021-10')
)
df.value = df.mask(mask, 0.0).value

使得df为

code       mapper       version  country  range    month              value
1       ABC321       ABC          Alpha     USA     High   2021-10              0.0
6       ABC321       ABC          Alpha     USA     High   2021-11              1.0
2       ABC321       PQS          Beta      IND     LOW    2021-10              0.0
3       ABC321       TRR          Delta     MEX     LOW    2021-10              1.0
4       ABC321       TRR          Delta     MEX     LOW    2021-11              3.0

现在我想添加蒙版值"8"它被更新为"0"给下个月，

expected_output =  code       mapper       version  country  range    month              value
1       ABC321       ABC          Alpha     USA     High   2021-10              0.0
6       ABC321       ABC          Alpha     USA     High   2021-11              9.0
2       ABC321       PQS          Beta      IND     LOW    2021-10              0.0
3       ABC321       TRR          Delta     MEX     LOW    2021-10              1.0
4       ABC321       TRR          Delta     MEX     LOW    2021-11              3.0

EDIT
There won't be duplicate rows

Answer 1

一个想法是将值转换为月周期，因此对于匹配前一个月或下一个月，仅使用+ 1或- 1:

df['month'] = pd.to_datetime(df['month']).dt.to_period('m')
mask = (
(df.mapper == 'ABC')
& (df.version == 'Alpha')
& (df.country == 'USA')
& (df.range == 'High')
& (df['month'] == '2021-10')
)
mask1 = (
(df.mapper == 'ABC')
& (df.version == 'Alpha')
& (df.country == 'USA')
& (df.range == 'High')
& (df['month'] - 1 == '2021-10')
)
df.loc[mask1, 'value'] += next(iter(df.loc[mask, 'value']), 0)
df.loc[mask, 'value'] = 0

print (df)
code mapper version country range    month  value
1  ABC321    ABC   Alpha     USA  High  2021-10    0.0
6  ABC321    ABC   Alpha     USA  High  2021-11    9.0
2  ABC321    PQS    Beta     IND   LOW  2021-10    0.0
3  ABC321    TRR   Delta     MEX   LOW  2021-10    1.0
4  ABC321    TRR   Delta     MEX   LOW  2021-11    3.0