Python 每个投资组合的价值加权投资组合回报

我有一个数据集，上面有一系列股票的日期、回报、投资组合和市值。我想在我的数据和每个投资组合中计算每月的价值加权市场回报

date        ret     portf   mkval
1982-03-31  0.02    3.0     2000
1982-04-30  0.05    2.0     500
1982-05-31  0.10    1.0     3000
1982-03-31  0.05    3.0     4000
1982-04-30  0.20    3.0     700
1982-05-31  0.02    2.0     2000
1982-05-31  0.08    1.0     5000

该数据应产生以下输出：

date         portf   equal_w_ret
1982-03-31   3.0     0.04
1982-04-30   2.0     0.05
1982-04-30   3.0     0.20
1982-05-31   1.0     0.0875
1982-05-31   2.0     0.02

这里，第一行计算为：（2000/（2000+4000））（1+0.02）+（4000/（2000+4000））（1+0.05）-1

提前谢谢

首先设置数据

data = { 'date' : ['1982-03-31','1982-04-30','1982-05-31','1982-03-31','1982-04-30','1982-05-31','1982-05-31'],
    'ret' : [0.02,0.05,0.10,0.05,0.20,0.02,0.08],
    'portf' : [3.0,2.0,1.0,3.0,3.0,2.0,1.0],
    'mkval' : [2000,500,3000,4000,700,2000,5000]}

现在将数据转换成数据帧并准备输出数据帧

df = pd.DataFrame(data)
dfout = pd.DataFrame()

for group, subdf in df.groupby(['date','portf']):
    subdf['wret'] = (subdf['mkval'] * ( 1 + subdf['ret']))/subdf['mkval'].sum()
    df2 = pd.DataFrame({ 'data' : [group[0]],'portf':[group[1]],'equal_w_ret':[subdf['wret'].sum() - 1]})
    dfout = dfout.append(df2)

有趣的一点！按日期和投资组合编号分组，然后按行进行计算。然后制作一个数据框，作为该行的摘要，并将其放入输出数据框中

df = pd.DataFrame(data)
dfout = pd.DataFrame()

for group, subdf in df.groupby(['date','portf']):
    subdf['wret'] = (subdf['mkval'] * ( 1 + subdf['ret']))/subdf['mkval'].sum()
    df2 = pd.DataFrame({ 'data' : [group[0]],'portf':[group[1]],'equal_w_ret':[subdf['wret'].sum() - 1]})
    dfout = dfout.append(df2)

这使得

    data        portf   equal_w_ret
0   1982-03-31  3.0     0.0400
0   1982-04-30  2.0     0.0500
0   1982-04-30  3.0     0.2000
0   1982-05-31  1.0     0.0875
0   1982-05-31  2.0     0.0200