Converting dates less than 1970 in python using pandas

Question

I want to convert dates of which are less than 1970 to pd.to_datetime format but am unable to do so as any date less than 1970 gets converted to the new millennium (>2000).

I want the dates to be converted as follows:

15/01/1967 to 1967-01-15
14/01/1968 to 1968-01-14
12/01/1969 to 1969-01-12

It rather gets converted as follows when I used pd.to_datetime().

15/01/1967 to 2067-01-15
14/01/1968 to 2068-01-14
12/01/1969 to 2069-01-12

I set the origin value in pd.to_datetime to be 1900 (as in the pandas documentation), but it doesn't recognise the term "origin".

Smells like trying to use a data type of a unix timestamp, where the date is the count of seconds since 1970 or so. In this data type there is no "before 1970". — Marek Puchalski, Jul 03 '20 at 20:47
The version on the notebook is 0.19.2. And I don''t think we can upgrade. This is a coursera notebook for an assignment. — Munakala Lohith, Jul 04 '20 at 03:44

score 0 · Accepted Answer · answered Jul 05 '20 at 12:42

0

works fine for me if I provide a specific format for pd.to_datetime:

import pandas as pd

s = pd.Series(["15/01/1967", "14/01/1968", "12/01/1969"])

pd.to_datetime(s, format='%d/%m/%Y')

# 0   1967-01-15
# 1   1968-01-14
# 2   1969-01-12
# dtype: datetime64[ns]

pandas 1.0.5 on Python 3.8.2 x64 / Windows 10.

answered Jul 05 '20 at 12:42

FObersteiner

22,500
8
42
72

Thank you. I tried to used origin as a feature in pd.to_datetime. – Munakala Lohith Jul 06 '20 at 13:27

Converting dates less than 1970 in python using pandas

1 Answers1