Question

我有一个看起来像这样的数据集

ID        TimeStamp        ScoreA ScoreB  Type       
A         20150908143000    345    316    New
B         20150908140300    400    480    New
B         20150908140600    NaN    120    Old
B         20150908143000  10743   8803    Old
C         20150908140100    600   1715    New
C         20150908140200    200   1062    Old
C         20150908141000    NaN    145    Old
C         20150908141500    418    NaN    Old
D         20150908143000    433     65    New

我希望结果看起来像这样

ID        TimeStamp       Score1  Score2  Type  FirstScore1  FirstScore2  
A         20150908143000    345    316    New
B         20150908140300    400    480    New
B         20150908140600    NaN    120    Old   400          480
B         20150908143000  10743   8803    Old   400          480
C         20150908140100    600   1715    New
C         20150908140200    200   1062    Old   600          1715
C         20150908141000    NaN    145    Old   600          1715
C         20150908141500    418    NaN    Old   600          1715
D         20150908143000    433     65    New

这样，只要“类型”等于“旧”，就会调用该特定“ ID”的最早的“ ScoreA”和“ ScoreB”，并将其分别放置在“ FirstScoreA”和“ FirstScoreB”中。

我已经能够提出使我获得最大价值的代码，但不是最早的。但是即使如此，我也不能仅将其限制为特定的ID，所以我倍受困扰。

有人可以帮我解决这个问题吗？

Answer 1

df['FirstScore1'] = np.where(df.Type=='New', df.ScoreA, np.nan)
df['FirstScore1'] = df.groupby('ID').FirstScore1.transform(lambda x: x.ffill())
df['FirstScore1'] = np.where(df.Type=='New', np.nan, df['FirstScore1'])
df['FirstScore2'] = np.where(df.Type=='New', df.ScoreB, np.nan)
df['FirstScore2'] = df.groupby('ID').FirstScore2.transform(lambda x: x.ffill())
df['FirstScore2'] = np.where(df.Type=='New', np.nan, df['FirstScore2'])


    ID  TimeStamp       ScoreA  ScoreB  Type    FirstScore1 FirstScore2
0   A   20150908143000  345.0   316.0   New     NaN         NaN
1   B   20150908140300  400.0   480.0   New     NaN         NaN
2   B   20150908140600  NaN 1   20.0    Old     400.0       480.0
3   B   20150908143000  10743.0 8803.0  Old     400.0       480.0
4   C   20150908140100  600.0   1715.0  New     NaN         NaN
5   C   20150908140200  200.0   1062.0  Old     600.0       1715.0
6   C   20150908141000  NaN     145.0   Old     600.0       1715.0
7   C   20150908141500  418.0   NaN     Old     600.0       1715.0
8   D   20150908143000  433.0   65.0    New     NaN         NaN

熊猫-召回特定群体的最早价值

1 个答案: