Question/bug: How to get rid of "0 days" in timedelta64[s]

[pllim]

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas. -- I can't tell if this is a feature or a bug.

• (optional) I have confirmed this bug exists on the master branch of pandas.

---

Note: Please read this guide detailing how to provide the necessary information for us to reproduce your bug.

Code Sample, a copy-pastable example

>>> import pandas as pd
>>> import numpy as np
>>> dt = pd.Series(np.array([1, 300], dtype='timedelta64[s]')) 
>>> df = pd.DataFrame({'dt': dt})

>>> pd.__version__
'1.0.5'
>>> df
        dt
0 00:00:01
1 00:05:00

>>> pd.__version__
'1.1.0'
>>> df
               dt
0 0 days 00:00:01
1 0 days 00:05:00

Problem description

"0 days" started to appear when DataFrame contains timedelta64[s].

Expected Output

See "Code Sample" for pandas 1.0.5.

Output of pd.show_versions()

See "Code Sample" for pandas 1.1.0.

Further readings

astropy/astropy#10608 and astropy/astropy#10609

[simonjayhawkins]

Thanks @pllim for the report. Not sure if this was an intentional change, so will mark as regression pending further investigation.

[simonjayhawkins]

It appears this was an intentional change, see tests changed in #33270

[pllim]

Wasn't expecting that in a minor release. Is there a way to get the old behavior back? If no, that's fine too, just want to know.

[simonjayhawkins]

@jbrockmendel

[jreback]

Wasn't expecting that in a minor release. Is there a way to get the old behavior back? If no, that's fine too, just want to know.

minor release can and do have api changes - pls read the release notes

this was for conformity of timedelta outputs

[pllim]

I did glance at the notes but I guess I missed it. Okie dokie then.

[pllim]

Thanks for digging into this, @simonjayhawkins !