Question

我有一个如下数据框：

df = pd.DataFrame({'Date': ['02/01/2019', '03/01/2019', '04/01/2019', '07/01/2019', '08/01/2019', '09/01/2019', '10/01/2019', '11/01/2019', '14/01/2019', '15/01/2019'],
                   'VOD': [3, 2.3, 2, 1.8, 2, 4, 5, 4, 3, 1],
                   'BBY': [0.9, 1, 1.2, 1, 1, 2.3, 2.4, 2.5, 3, 2.9],
                   'STJ': [4, 4.2, 4.3, 4.4, 3.5, 3, 2, 1, 1.2, 2],
                   'RBS': [0.5, 0.6, 0.7, 0.6, 1, 1.2, 1.3, 1.4, 1.5, 2]})

从此数据框中，我可以按列对每一行进行排序，如下所示：

df1 = df.rank(1, ascending=False, method='first')

我正在尝试将1分配给排名最高的两个（在第一行中是VOD和STJ），将0分配给其他。

我的目标是建立一个如下表：

result = pd.DataFrame({'Date': ['02/01/2019', '03/01/2019', '04/01/2019', '07/01/2019', '08/01/2019', '09/01/2019', '10/01/2019', '11/01/2019', '14/01/2019', '15/01/2019'],
                       'VOD': [1, 1, 1, 1, 1, 1, 1, 1, 1, 0],
                       'BBY': [0,0,0,0,0,0,1,1,1,1],
                       'STJ': [1,1,1,1,1,1,0,0,0,1],
                       'RBS': [0,0,0,0,0,0,0,0,0,0]})

我认为if语句会起作用，但不能使rank（）起作用。想法大受欢迎。

Answer 1

使用DataFrame.isin进行True/False到1/0映射的强制转换为整数：

cols = ['VOD','BBY','STJ','RBS']
df[cols] = df[cols].rank(axis=1, ascending=False, method='first').isin([1,2]).astype(int)

或使用numpy.where：

df[cols] = np.where(df[cols].rank(axis=1, ascending=False, method='first').isin([1,2]), 1, 0)

print (df)
         Date  VOD  BBY  STJ  RBS
0  02/01/2019    1    0    1    0
1  03/01/2019    1    0    1    0
2  04/01/2019    1    0    1    0
3  07/01/2019    1    0    1    0
4  08/01/2019    1    0    1    0
5  09/01/2019    1    0    1    0
6  10/01/2019    1    1    0    0
7  11/01/2019    1    1    0    0
8  14/01/2019    1    1    0    0
9  15/01/2019    0    1    1    0

Answer 2

import pandas as pd

df = pd.DataFrame({'Date': ['02/01/2019', '03/01/2019', '04/01/2019', '07/01/2019', '08/01/2019', '09/01/2019', '10/01/2019', '11/01/2019', '14/01/2019', '15/01/2019'],
                   'VOD': [3, 2.3, 2, 1.8, 2, 4, 5, 4, 3, 1],
                   'BBY': [0.9, 1, 1.2, 1, 1, 2.3, 2.4, 2.5, 3, 2.9],
                   'STJ': [4, 4.2, 4.3, 4.4, 3.5, 3, 2, 1, 1.2, 2],
                   'RBS': [0.5, 0.6, 0.7, 0.6, 1, 1.2, 1.3, 1.4, 1.5, 2]})


ranked_cols = ['VOD','BBY','STJ','RBS']
ranked = df[ranked_cols].rank(axis=1, ascending=False, method='first')

def allocate_ones(x):
    if x in (1, 2):  # top 2 ranked
        return 1
    else:
        return 0

allocated = ranked.applymap(allocate_ones)

现在重新附加日期列：

allocated['Date'] = df['Date']

输出：

   VOD  BBY  STJ  RBS        Date
0    1    0    1    0  02/01/2019
1    1    0    1    0  03/01/2019
2    1    0    1    0  04/01/2019
3    1    0    1    0  07/01/2019
4    1    0    1    0  08/01/2019
5    1    0    1    0  09/01/2019
6    1    1    0    0  10/01/2019
7    1    1    0    0  11/01/2019
8    1    1    0    0  14/01/2019
9    0    1    1    0  15/01/2019

熊猫-如果排名前2位，则排名1

2 个答案: