0

In a pandas DataFrame I have weird datetime format like so: 

0    201913907050435
1    201913908520126
2    201914004163647
3    201914019315651
4    201914019320917
Name: DATETIME, dtype: object

What I know is, that it's Year followed by day of the year. I guess the number after that is the time in milliseconds or nanoseconds.

I want to transform the type to a pandas datetime object. So far I have:

pd.to_datetime(df.DATETIME, format='%Y%j%f')

which gives me:

0   2019-05-19 00:00:00.070504350
1   2019-05-19 00:00:00.085201260
2   2019-05-20 00:00:00.041636470
3   2019-05-20 00:00:00.193156510
4   2019-05-20 00:00:00.193209170
Name: DATETIME, dtype: datetime64[ns]

The problem is that HH:MM:SS is always 0 and not filled in correctly. What datetime code do I need instead of %f that can represent milliseconds (or whatever this number is)?

I could not find it in this table.

Thank you!

Sebastian N
  • 131
  • 2
  • 9

1 Answers1

1
### try this
print(pd.to_datetime('201913907050435', format="%Y%j%H%M%S%f"))
##output:
2019-05-19 07:05:04.350000
Sebastian N
  • 131
  • 2
  • 9
vrana95
  • 511
  • 2
  • 10
  • If you would have read the question carefully you would have seen that it's day-of-year (code %j) at second place. But with your answer I think I got to the correct solution. Please accept my edits then I will accept your answer. – Sebastian N Jun 03 '19 at 11:45
  • i have accepted your edit. I missed little part. Thnx for edit – vrana95 Jun 03 '19 at 11:49