python - 使用基于另一行的条件删除组行

Question

我有一个电影数据框，如下所示:

Index  Movie                   Rotten Tomato Score  Director             Ranking
1      Batman Rises            87                   Christopher Nolan    1
2      Interstellar            73                   Christopher Nolan    2
3      Legally Blonde          71                   Robert Luketic       1
4      The Ugly Truth          14                   Robert Luketic       2
5      (500) Days of Summer    85                   Marc Webb            1
6      The Amazing Spider-Man  71                   Marc Webb            2
7      Wide Awake              45                   M N Shyamalan        1
8      The Last Airbender       5                   M N Shyamalan        2

我制作了一个排名栏，根据烂番茄百分比显示排名顺序。这是按电影导演分组的。

我想做的是使用排名列，如果排名靠前的电影得分高于某个点(50%)，则删除排名较低的电影。例如，对于 Marc Webb，我唯一想为他放映的电影是《夏日的 500 天》，但是我想为 M N Shyamalan 放映这两部电影。理想的表格如下所示:

Index  Movie                 Rotten Tomato Score  Director             Ranking  
    1  Batman Rises          87                   Christopher Nolan    1      
    3  Legally Blonde        71                   Robert Luketic       1
    5  (500) Days of Summer  85                   Marc Webb            1
    7  Wide Awake            45                   M N Shyamalan        1
    8  The Last Airbender    5                    M N Shyamalan        2

我尝试过:

movie_names = movie_names.groupby('Movie').filter(lambda g: (g.score <= 0.5).any())

然而，这删除了 ​​M N Shyamalan 的两部电影。

有人知道该怎么做吗？任何帮助将不胜感激!

Answer 1

代码

# is movie score > 50?
m = df['Rotten Tomato Score'] > 50

# Does director has at least one movie with score > 50?
cond1 = m.groupby(df['Director']).transform('any')

# flag the duplicate rows and keep the movie with highest score
cond2 = df.sort_values('Rotten Tomato Score').duplicated('Director', keep='last')

# Drop rows when cond1 and cond2 is met
df[~(cond1 & cond2)]

结果

   Index                 Movie  Rotten Tomato Score           Director  Ranking
0      1          Batman Rises                   87  Christopher Nolan        1
2      3        Legally Blonde                   71     Robert Luketic        1
4      5  (500) Days of Summer                   85          Marc Webb        1
6      7            Wide Awake                   45      M N Shyamalan        1
7      8    The Last Airbender                    5      M N Shyamalan        2