平均分配一個月的量爲周

Question

有兩個數據幀：

1）DF

Month   Qty 
    ------------------- 
0 2017-10-31  100 
1 2017-11-30  200 

2）周

Week 
    ---------- 
0 2017-10-01 
1 2017-10-08 
2 2017-10-15 
3 2017-10-22 
4 2017-10-29 
5 2017-11-05 
6 2017-11-12 
7 2017-11-19 
8 2017-11-26 

我怎麼能同樣跨周的時間才能發佈每月的數量以下輸出：

Week   Qty 
    ------------------- 
0 2017-10-01  20 
1 2017-10-08  20 
2 2017-10-15  20 
3 2017-10-22  20 
4 2017-10-29  20 
5 2017-11-05  50 
6 2017-11-12  50 
7 2017-11-19  50 
8 2017-11-26  50 

Answer 1

轉換datetimes到month期間由to_period在兩個df。

df1['Month'] = pd.to_datetime(df1['Month']).dt.to_period('m') 
df2['Week'] = pd.to_datetime(df2['Week']) 

s = df1.set_index('Month')['Qty'] 
a = df2['Week'].dt.to_period('m') 
df2['Qty'] = a.map(s)/df2.index.to_series().groupby(a).transform('size') 
print (df2) 
     Week Qty 
0 2017-10-01 20.0 
1 2017-10-08 20.0 
2 2017-10-15 20.0 
3 2017-10-22 20.0 
4 2017-10-29 20.0 
5 2017-11-05 50.0 
6 2017-11-12 50.0 
7 2017-11-19 50.0 
8 2017-11-26 50.0 

詳細說明：：

然後通過transform創建的每個週期的長度創建Series在地圖柱Week和除法

print (a.map(s)) 
0 100 
1 100 
2 100 
3 100 
4 100 
5 200 
6 200 
7 200 
8 200 
Name: Week, dtype: int64 

print (df2.index.to_series().groupby(a).transform('size')) 
0 5 
1 5 
2 5 
3 5 
4 5 
5 4 
6 4 
7 4 
8 4 
dtype: int64