Pandas：使用文本旋转数据框并合并列

Question

我正在与 Python 和 Pandas 合作，并且有一个 table 像这样：

      Name        Team    Fixture   Line-up     Min IN   Min Out
0     Player 1    RAY     J1        Starting             68
1     Player 2    RAY     J1        Bench       74       
2     Player 3    RSO     J2        Starting             45
3     Player 4    RSO     J2        Bench       45

我需要旋转 table 使 'Fixture' 的行成为包含 'Line-up' 的文本 + Min IN 和 OUT 的新列。那么结果应该是这样的：

      Name        Team    J1                J2
0     Player 1    RAY     Starting - 68
1     Player 2    RAY     Bench - 74      
2     Player 3    RSO                       Starting - 45
3     Player 4    RSO                       Bench - 45

有什么办法可以做到吗？提前致谢！

Answer 1

您可以通过包含最小值修改 Line-up 列，然后 pivot:

out = (df.assign(**{'Line-up': df['Line-up'] + ' - ' + 
                    df.filter(like='Min').bfill(axis=1).iloc[:,0].astype(int).astype(str)})
       .pivot(['Name','Team'], 'Fixture','Line-up').rename_axis(columns=None).reset_index())

输出：

       Name Team             J1             J2
0  Player 1  RAY  Starting - 68            NaN
1  Player 2  RAY     Bench - 74            NaN
2  Player 3  RSO            NaN  Starting - 45
3  Player 4  RSO            NaN     Bench - 45

N.B。这假定 Min 列中的空 space 是 NaN 值。如果它们实际上是空 space '' ，那么您可以先将它们转换为 NaN 值。所以喜欢：

out = (df.assign(**{'Line-up': df['Line-up'] + ' - ' + 
                    df.filter(like='Min').replace('', pd.NA).bfill(axis=1).iloc[:,0].astype(int).astype(str)})
#                               here -->  ^^^^^^^^^^^^
       .pivot(['Name','Team'], 'Fixture','Line-up').rename_axis(columns=None).reset_index())

Answer 2

另一个版本：

df = (
    df.set_index(["Name", "Team", "Fixture"])
    .apply(lambda x: " - ".join(x[x != ""]), axis=1)
    .unstack(level=2)
    .reset_index()
)
df.columns.name = ""

打印：

       Name Team             J1             J2
0  Player 1  RAY  Starting - 68            NaN
1  Player 2  RAY     Bench - 74            NaN
2  Player 3  RSO            NaN  Starting - 45
3  Player 4  RSO            NaN     Bench - 45

Pandas：使用文本旋转数据框并合并列

Pandas: Pivot dataframe with text and combine columns

python

pivot-table

pandas