python - 使用 pandas size() 函数包含零计数的项目？

Question

我有一个由一些 Jira 问题组成的数据框，我试图按周和状态对其进行排序，并获取每周每个状态的项目数量。例如，我的数据帧的两周目前如下所示:

2019-11-04     Authorize Work      4
               Await Work          1
               Check Work          4
               Closed              4
               Confirm Work        3
               Do Work             3
2019-11-11     Authorize Work      6
               Do Work             2

我已经通过以下内容达到了这一点:

# Remove the time portion of the date
df['creation_date'] = df['creation_date'].dt.date
# Set the date to be a week long delta
df['creation_date'] = pd.to_datetime(df['creation_date']) - pd.to_timedelta(7, unit='d')
# Sort together by creation date within the week and the status, and do a count
endf = df.groupby([pd.Grouper(key="creation_date", freq="W-MON"), "status"]).size()

您会注意到第二周只有两种状态，而第一周有六个状态。这是因为第二周内缺失状态的 jira 问题为零。有没有办法让大小函数包含计数为零的缺失状态，以便每周内的数据都是同一组状态？

Answer 1

您可以尝试unstack和stack:

enddf.unstack(level=-1, fill_value=0).stack()

输出:

                           2
0          1                
2019-11-04 Authorize Work  4
           Await Work      1
           Check Work      4
           Closed          4
           Confirm Work    3
           Do Work         3
2019-11-11 Authorize Work  6
           Await Work      0
           Check Work      0
           Closed          0
           Confirm Work    0
           Do Work         2