Extracting/deleting rows from time series without using index information

Question

I have a simple problem. I want to get a subset of time series with a certain condition that is not dependent on time index. I have a very huge dataset, I am just giving a small example to make my problem understandable.

	row_num	marks
2016-01-01	1	99
2016-01-02	2	98
2016-01-01	3	95
2016-01-01	4	90
2016-01-02	5	40
2016-01-03	6	80
2016-01-04	7	20

but my problem is when I try to drop, it always drops by index and delete all index of 2016-01-01 and 2016-01-02. I can not manually extract such a subset because data size is very huge and there are so many duplicate indices. How to solve this problem?

Answer 1

Try using the following:

df[~df['row_num'].isin([1,5])]

Output:

              row_num  marks
  2016-01-02        2     98
  2016-01-01        3     95
  2016-01-01        4     90
  2016-01-03        6     80
  2016-01-04        7     20