在Python中每个月末/月初将运行计数重置为零

Question

我正在创建一个索赔的运行计数，并希望在每个月初将计数重置为0。

我现在有这些数据帧:

DF1:

<表类> 日期 X Y tbody><<tr>2021-04-25442021-04-26002021-04-27002021-04-2856532021-04-29002021-04-30102021-05-01002021-05-02502021-05-03552021-05-0400

尝试:

Groupby (df[日期].dt.Month) .agg (cumsum)

这应该可以为您工作:

df = pd.DataFrame([
('2021-04-25', 4, 4),
('2021-04-26', 0, 0),
('2021-04-27', 0, 0),
('2021-04-28', 56, 53),
('2021-04-29', 0, 0),
('2021-04-30', 1, 0),
('2021-05-01', 0, 0),
('2021-05-02', 5, 0),
('2021-05-03', 5, 5),
('2021-05-04', 0, 0),
], columns=['Date', 'X', 'Y',])
df['Date'] = pd.to_datetime(df['Date'])
# Create Cumulative Sums for X 
df['X'] = (df['X'] - df['Y'].shift(fill_value=0)).cumsum()
result = pd.concat([
group.assign(X=(
group['X'] - group.loc[group['Date'].dt.is_month_start, 'X'].values[0]
if len(group.loc[group['Date'].dt.is_month_start]) > 0
else group['X']
))
for month, group in df.groupby(df['Date'].dt.month)
])

输出:

Date   X   Y
0 2021-04-25   4   4
1 2021-04-26   0   0
2 2021-04-27   0   0
3 2021-04-28  56  53
4 2021-04-29   3   0
5 2021-04-30   4   0
6 2021-05-01   0   0
7 2021-05-02   5   0
8 2021-05-03  10   5
9 2021-05-04   5   0