python - 如何计算 Pandas 每个时期的增加量

Question

我有一个 df，其中包含每个时间段的 JIRA 票证状态摘要，它包含“打开”、“关闭”和“其他”的计数。我想看到随着时间的推移，超额门票数量的增加。

period                              Status  Counts
No. 1 Apr 06 2019 to Apr 12 2019    CLOSE   1026
No. 1 Apr 06 2019 to Apr 12 2019    OPEN    2914
No. 1 Apr 06 2019 to Apr 12 2019    OTHER   264
No. 2 Mar 30 2019 to Apr 05 2019    CLOSE   1307
No. 2 Mar 30 2019 to Apr 05 2019    OPEN    2212
No. 2 Mar 30 2019 to Apr 05 2019    OTHER   256 

对于第 1 期，OPEN 状态的计数已从 2212(第 2 期)增加到2914，因此为第 1 时段添加了 702 张票。如何添加显示的额外列。

period                              Status  Counts   Added
No. 1 Apr 06 2019 to Apr 12 2019    CLOSE   1026     702 (2914-2212)
No. 1 Apr 06 2019 to Apr 12 2019    OPEN    2914     702 
No. 1 Apr 06 2019 to Apr 12 2019    OTHER   264      702 
No. 2 Mar 30 2019 to Apr 05 2019    CLOSE   1307     (2212 minus  xxx)
No. 2 Mar 30 2019 to Apr 05 2019    OPEN    2212     (2212 minus  xxx)
No. 2 Mar 30 2019 to Apr 05 2019    OTHER   256      (2212 minus  xxx)

Answer 1

您可以在OPEN中找到差异，然后使用transform('first')将这些值放回框架中。

u = df.assign(Added=df.loc[df.Status.eq('OPEN'), 'Counts'].diff(-1))

u.assign(Added=u.groupby('period')['Added'].transform('first'))

                             period Status  Counts  Added
0  No. 1 Apr 06 2019 to Apr 12 2019  CLOSE    1026  702.0
1  No. 1 Apr 06 2019 to Apr 12 2019   OPEN    2914  702.0
2  No. 1 Apr 06 2019 to Apr 12 2019  OTHER     264  702.0
3  No. 2 Mar 30 2019 to Apr 05 2019  CLOSE    1307    NaN
4  No. 2 Mar 30 2019 to Apr 05 2019   OPEN    2212    NaN
5  No. 2 Mar 30 2019 to Apr 05 2019  OTHER     256    NaN