熊猫 timedelta 和过滤负值

Question

I have a pandas dataframe with column time_taken as a timedelta64[ns]. Here's a example of the values :

idle_time["time_taken"]

0                   NaT
1       00:01:41.648602
2       00:01:40.224034
3       00:01:44.029010
4       00:01:35.649479
              ...      
11608   00:33:18.597127
11609   00:03:09.414821
11610   00:04:43.362200
11611   00:25:36.113369
11612   00:06:27.766094
Name: time_taken, Length: 11613, dtype: timedelta64[ns]

I want to filter for negative values, i know there is one because running the statement idle_time["time_taken"].min() shows a negative. However i want to filter all rows with a negative.

I've tried idle_time[idle_time["time_taken"] > 0] but it shows

Invalid comparison between dtype=timedelta64[ns] and int

Any help is appreciated.

Answer 1

Convert timedeltas to some numbers, eg by Series.dt.total_seconds :

idle_time[idle_time["time_taken"].dt.total_seconds() > 0]

Or compare by Timedelta :

idle_time[idle_time["time_taken"] > pd.Timedelta(0)]