[英]Pandas Dataframe: Find the index of maximum value for each column using idxmax()
[英]find the maximum value for each streak of numbers in another column in pandas
我有一個像這樣的數據框:
df = pd.DataFrame({'dir': [1,1,1,1,0,0,1,1,1,0], 'price':np.random.randint(100,200,10)})
dir price
0 1 100
1 1 150
2 1 190
3 1 194
4 0 152
5 0 151
6 1 131
7 1 168
8 1 112
9 0 193
並且我想要一個新列來顯示最高價格,只要dir為1,並在dir為0時重置。我想要的結果如下所示:
dir price max
0 1 100 194
1 1 150 194
2 1 190 194
3 1 194 194
4 0 152 NaN
5 0 151 NaN
6 1 131 168
7 1 168 168
8 1 112 168
9 0 193 NaN
transform
max
transform
用於已過濾的行:
#get unique groups for consecutive values
g = df['dir'].ne(df['dir'].shift()).cumsum()
#filter only 1
m = df['dir'] == 1
df['max'] = df[m].groupby(g)['price'].transform('max')
print (df)
dir price max
0 1 100 194.0
1 1 150 194.0
2 1 190 194.0
3 1 194 194.0
4 0 152 NaN
5 0 151 NaN
6 1 131 168.0
7 1 168 168.0
8 1 112 168.0
9 0 193 NaN
聲明:本站的技術帖子網頁,遵循CC BY-SA 4.0協議,如果您需要轉載,請注明本站網址或者原文地址。任何問題請咨詢:yoyou2525@163.com.