如何根据ID在新列中镜像值

Question

我拥有的数据集包含NHL播放器数据，看起来像这样（简化）：

 teamNAME   playerID        gameID     metric       Won/Lost
  CAP       8473345           20          10           1
  WILD      8475467           20          10           0     
  NY        8471345           21          10           1
  STARS     8475756           21          10           0

我想做的是在所有行中都有对手teamNAME，这样看起来像这样：

 teamNAME   playerID        gameID     metric       Won/Lost    Opponent
  CAP       8473345           20          10           1         WILD
  WILD      8475467           20          10           0          CAP    
  NY        8471345           21          10           1        STARS
  STARS     8475756           21          10           0           NY

您如何解决这个问题？

Answer 1

您可以GroupBy gameID并结合使用transform和lambda函数来旋转值：

df['Opponent'] = df.groupby('gameID').teamNAME.transform(lambda x: x[::-1])

   teamNAME  playerID  gameID  metric  Won/Lost Opponent
0      CAP   8473345      20      10         1     WILD
1     WILD   8475467      20      10         0      CAP
2       NY   8471345      21      10         1    STARS
3    STARS   8475756      21      10         0       NY

Answer 2

从SQL角度解决此问题，您可以对gameID执行交叉联接，然后对teamNAME进行过滤：

(df.merge(df[['teamNAME', 'gameID']], on='gameID', how='left')
   .query('teamNAME_x != teamNAME_y')
   .rename({'teamNAME_x': 'teamNAME', 'teamNAME_y': 'Opponent'}, axis=1))

  teamNAME  playerID  gameID  metric  Won/Lost Opponent
1      CAP   8473345      20      10         1     WILD
2     WILD   8475467      20      10         0      CAP
5       NY   8471345      21      10         1    STARS
6    STARS   8475756      21      10         0       NY

Answer 3

将groupby与itertools.chain组合：

from itertools import chain

grouper = df.groupby('gameID')['teamNAME']
df['Opponent'] = list(chain.from_iterable(x.values[::-1] for _, x in grouper))

print(df)

  teamNAME  playerID  gameID  metric  Won/Lost Opponent
0      CAP   8473345      20      10         1     WILD
1     WILD   8475467      20      10         0      CAP
2       NY   8471345      21      10         1    STARS
3    STARS   8475756      21      10         0       NY