Python 熊猫：基于多级首次出现合并数据帧

我试图根据一对条件（

month

和

num

）的首次出现，将一个大数据帧与一个小数据帧合并

我已经拼凑出了可行的代码（底部是实际的/期望的输出），但它似乎可以更有效

我的问题是——我是否错过了一个更简单的方法

设置：

import pandas as pd

m = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2]
n = [1,1,1,20,20,300,300,20,20,1,1,1,20,300,20,1,1,1,20,20,300,300,300,20,20,1,1]
df = pd.DataFrame({'month':m, 'num':n, 'x':0})

m2 = [1,1,1,2,2,2]
n2 = [1,20,300,1,20,300]
s2 = [11,222,3333,44,555,6666]
df2 = pd.DataFrame({'month':m2, 'num':n2, 'sum':s2})

dfx = pd.DataFrame(df.groupby(['month','num'])['x'].idxmax())
dfx.rename(columns = {'x':'find'}, inplace = True)

df2.set_index(['month','num'], inplace = True)
df2 = pd.merge(df2, dfx, left_index = True, right_index = True)

df = df.merge(df2, left_index = True, right_on = 'find', how = 'left')
df = df.drop(['find','x'], axis = 1).reset_index(drop = True).fillna(0)

    month  num     sum
0       1    1    11.0
1       1    1     0.0
2       1    1     0.0
3       1   20   222.0
4       1   20     0.0
5       1  300  3333.0
6       1  300     0.0
7       1   20     0.0
8       1   20     0.0
9       1    1     0.0
10      1    1     0.0
11      1    1     0.0
12      1   20     0.0
13      1  300     0.0
14      1   20     0.0
15      1    1     0.0
16      2    1    44.0
17      2    1     0.0
18      2   20   555.0
19      2   20     0.0
20      2  300  6666.0
21      2  300     0.0
22      2  300     0.0
23      2   20     0.0
24      2   20     0.0
25      2    1     0.0
26      2    1     0.0

当前代码：

import pandas as pd

m = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2]
n = [1,1,1,20,20,300,300,20,20,1,1,1,20,300,20,1,1,1,20,20,300,300,300,20,20,1,1]
df = pd.DataFrame({'month':m, 'num':n, 'x':0})

m2 = [1,1,1,2,2,2]
n2 = [1,20,300,1,20,300]
s2 = [11,222,3333,44,555,6666]
df2 = pd.DataFrame({'month':m2, 'num':n2, 'sum':s2})

dfx = pd.DataFrame(df.groupby(['month','num'])['x'].idxmax())
dfx.rename(columns = {'x':'find'}, inplace = True)

df2.set_index(['month','num'], inplace = True)
df2 = pd.merge(df2, dfx, left_index = True, right_index = True)

df = df.merge(df2, left_index = True, right_on = 'find', how = 'left')
df = df.drop(['find','x'], axis = 1).reset_index(drop = True).fillna(0)

    month  num     sum
0       1    1    11.0
1       1    1     0.0
2       1    1     0.0
3       1   20   222.0
4       1   20     0.0
5       1  300  3333.0
6       1  300     0.0
7       1   20     0.0
8       1   20     0.0
9       1    1     0.0
10      1    1     0.0
11      1    1     0.0
12      1   20     0.0
13      1  300     0.0
14      1   20     0.0
15      1    1     0.0
16      2    1    44.0
17      2    1     0.0
18      2   20   555.0
19      2   20     0.0
20      2  300  6666.0
21      2  300     0.0
22      2  300     0.0
23      2   20     0.0
24      2   20     0.0
25      2    1     0.0
26      2    1     0.0

输出：

import pandas as pd

m = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2]
n = [1,1,1,20,20,300,300,20,20,1,1,1,20,300,20,1,1,1,20,20,300,300,300,20,20,1,1]
df = pd.DataFrame({'month':m, 'num':n, 'x':0})

m2 = [1,1,1,2,2,2]
n2 = [1,20,300,1,20,300]
s2 = [11,222,3333,44,555,6666]
df2 = pd.DataFrame({'month':m2, 'num':n2, 'sum':s2})

dfx = pd.DataFrame(df.groupby(['month','num'])['x'].idxmax())
dfx.rename(columns = {'x':'find'}, inplace = True)

df2.set_index(['month','num'], inplace = True)
df2 = pd.merge(df2, dfx, left_index = True, right_index = True)

df = df.merge(df2, left_index = True, right_on = 'find', how = 'left')
df = df.drop(['find','x'], axis = 1).reset_index(drop = True).fillna(0)

    month  num     sum
0       1    1    11.0
1       1    1     0.0
2       1    1     0.0
3       1   20   222.0
4       1   20     0.0
5       1  300  3333.0
6       1  300     0.0
7       1   20     0.0
8       1   20     0.0
9       1    1     0.0
10      1    1     0.0
11      1    1     0.0
12      1   20     0.0
13      1  300     0.0
14      1   20     0.0
15      1    1     0.0
16      2    1    44.0
17      2    1     0.0
18      2   20   555.0
19      2   20     0.0
20      2  300  6666.0
21      2  300     0.0
22      2  300     0.0
23      2   20     0.0
24      2   20     0.0
25      2    1     0.0
26      2    1     0.0

---

如果我理解正确，您可以在两个数据帧之间执行常规操作，然后合并并将非首次出现的事件归零：

df3 = df.merge(df2, how='left', on=['month', 'num'])
df3.loc[df3.duplicated(subset=['month', 'num']), 'sum'] = 0

结果输出：

   month  num   sum
0       1    1    11
1       1    1     0
2       1    1     0
3       1   20   222
4       1   20     0
5       1  300  3333
6       1  300     0
7       1   20     0
8       1   20     0
9       1    1     0
10      1    1     0
11      1    1     0
12      1   20     0
13      1  300     0
14      1   20     0
15      1    1     0
16      2    1    44
17      2    1     0
18      2   20   555
19      2   20     0
20      2  300  6666
21      2  300     0
22      2  300     0
23      2   20     0
24      2   20     0
25      2    1     0
26      2    1     0

---

很不错的！我现在可以继续了：-）啊，因为默认情况下，

.duplicated

保留第一次出现。他妈的天才。非常感谢。