我有这个数据框df
delta_clean delta_time new_index
2017-05-24 11:59:29 -2.837001e-01 1.472778 0
2017-05-24 13:02:18 -2.546754e-02 1.046944 0
2017-05-24 14:05:06 -1.039444e-01 1.046667 0
2017-05-24 15:29:12 5.573439e+00 1.401667 0
2017-05-24 16:32:02 -1.567387e-01 1.047222 0
2017-05-24 17:34:52 -8.079458e-02 1.047222 0
2017-05-24 18:37:41 -5.220998e-02 1.046944 0
2017-05-24 19:40:31 -2.302334e-02 1.047222 0
2017-05-24 20:43:20 -4.382400e-02 1.046944 0
2017-05-24 21:46:10 -1.675438e-02 1.047222 0
2017-05-24 22:49:01 -5.994632e-02 1.047500 0
2017-05-24 23:51:53 -4.845164e-02 1.047778 0
2017-05-24 23:59:59 -2.226252e-02 0.135000 0
2017-05-25 00:00:01 -9.161529e-05 0.000556 0
2017-05-25 00:54:45 -1.504323e-01 0.912222 0
2017-05-25 01:57:38 -5.539247e-02 1.048056 0
2017-05-25 06:34:57 -2.389934e-01 4.621944 0
2017-05-25 07:37:48 1.037526e-01 1.047500 0
2017-05-25 08:40:38 -2.081650e-01 1.047222 0
2017-05-25 09:43:27 -5.335963e-02 1.046944 0
2017-05-25 10:46:16 -8.687169e-02 1.046944 0
2017-05-25 11:49:05 -2.209528e-01 1.046944 0
2017-05-25 12:51:54 -4.416145e-02 1.046944 0
2017-05-25 13:54:43 -3.051953e-02 1.046944 0
2017-05-25 14:57:31 -4.792480e-02 1.046667 0
2017-05-25 16:00:20 -1.288275e-02 1.046944 0
2017-05-25 17:03:09 3.362946e-02 1.046944 0
2017-05-25 18:05:59 -6.533723e-02 1.047222 0
2017-05-25 19:08:49 -6.509860e-02 1.047222 0
我每次发现此条件df['delta_clean']/df['delta_time']>2时都需要增加新索引的值
我该怎么办？
答案 0 :(得分：1)
要累积增加序列号，可以使用pd.Series.cumsum：
df['new_index'] = (df['delta_clean'] / df['delta_time'] > 2).cumsum()
之所以可行，是因为在进行数值计算时，布尔系列（例如True）中的(df['delta_clean'] / df['delta_time'] > 2)个值被解释为1。类似于常规Python，其中bool是int的子类。