英文:
pandas window by groupby function
问题
我有一个pd数据框：
pd.DataFrame({'day': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],'A': [1, 1, 1, 0, 0, 0, 1, 1, 0, 0],'B': [7, 8, 4, 3, 5, 6, 3, 1, 4, 5],})
我想要创建一个滑动窗口，只要列的值为1，就找到最大值，一旦值变为0，就找到这些0的最大值。以下是期望的输出：
pd.DataFrame({'day': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],'A': [1, 1, 1, 0, 0, 0, 1, 1, 0, 0],'B': [7, 8, 4, 3, 5, 6, 3, 1, 4, 5],'Output': [8, 8, 8, 6, 6, 6, 3, 3, 5, 5 ]})
英文:
I have a pd dataframe
pd.DataFrame({'day': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],'A': [1, 1, 1, 0, 0,0, 1, 1, 0, 0],'B': [7, 8, 4, 3, 5, 6, 3, 1, 4, 5],})
I want to do a sliding window that as long as the value of a column is 1 find the max, as soon as the value changes to 0 find the max for those 0. here is a desired output:
pd.DataFrame({'day': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],'A': [1, 1, 1, 0, 0,0, 1, 1, 0, 0],'B': [7, 8, 4, 3, 5, 6, 3, 1, 4, 5],'Output': [8, 8, 8, 6, 6, 6, 3, 3, 5, 5 ]})
答案1
得分: 0
使用A的移位值和Series.cumsum来创建连续的分组，然后传递给GroupBy.transform，使用max函数：
df['Output'] = df.groupby(df.A.ne(df.A.shift()).cumsum())['B'].transform('max')print(df)day A B Output0 1 1 7 81 2 1 8 82 3 1 4 83 4 0 3 64 5 0 5 65 6 0 6 66 7 1 3 37 8 1 1 38 9 0 4 59 10 0 5 5
英文:
Create consecutive groups by A by compare with shifted values and Series.cumsum, then pass to GroupBy.transform with max:
df['Output'] = df.groupby(df.A.ne(df.A.shift()).cumsum())['B'].transform('max')print (df)day A B Output0 1 1 7 81 2 1 8 82 3 1 4 83 4 0 3 64 5 0 5 65 6 0 6 66 7 1 3 37 8 1 1 38 9 0 4 59 10 0 5 5
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论