How can I take one nanosecond away from a datetime timestamp, and have the final form in datetime timestamp?

Question

I have a datetime timestamp of 2022-10-11 06:00:00.

I am trying to take away ONE nanosecond and have it back in datetime format. So the desired output is:

`2022-10-11 05:59:59.999999999`

I have followed a few functions and conversions on stack overflow, but most don't get the precision correct. They convert the datetime into nanoseconds, take one away then convert it back, but then one hour is somehow added because of precision (or lack of precision)

Any help would be great!

That's not one nanosecond less, that's 10 microseconds less. — jprebys, Jan 13 '23 at 20:50
Please [edit] your question and post [what you have tried so far](https://meta.stackoverflow.com/q/261592). — MattDMo, Jan 13 '23 at 20:51
Datetime doesn't go down to nanoseconds, microseconds is as low as it goes. — Mark Ransom, Jan 13 '23 at 20:52
It's still not one nanosecond less, that would be `2022-10-11 05:59:59.999999999` — jprebys, Jan 13 '23 at 20:53
Does this answer your question? [Get POSIX/Unix time in seconds and nanoseconds in Python?](https://stackoverflow.com/questions/2394485/get-posix-unix-time-in-seconds-and-nanoseconds-in-python) — sahasrara62, Jan 13 '23 at 20:55
How does it work, you pass in the time as an argument into`time.time_ns()`? — , Jan 13 '23 at 20:57
OK I see what it does, but how does it help me?? - I'm trying to take one nanosecond off. — , Jan 13 '23 at 20:57
It doesn't help you. The combination of datetime and nanoseconds simply can't be done. — Mark Ransom, Jan 13 '23 at 21:19

score 0 · Accepted Answer · answered Jan 14 '23 at 01:39

datetime.datetime only supports till microsecond precision. If you need nanosecond precision, use pandas.Timestamp instead.

import pandas as pd
import time

now_in_ns = pd.Timestamp(time.time_ns(), unit='ns')
one_ns_later = now_in_ns + pd.Timedelta('1ns')

print(now_in_ns, one_ns_later)

So, if you are going to be working with nanoseconds, either extend datetime.datetime to support nanosecond or familiarize yourself with the class: https://pandas.pydata.org/docs/reference/api/pandas.Timestamp.html

How can I take one nanosecond away from a datetime timestamp, and have the final form in datetime timestamp?

1 Answers1