Naturally sorting Pandas DataFrame

Using sort_values for pandas >= 1.1.0

With the new key argument in DataFrame.sort_values, since pandas 1.1.0, we can directly sort a column without setting it as an index using natsort.natsort_keygen:

df = pd.DataFrame({
    "time": ['0hr', '128hr', '72hr', '48hr', '96hr'],
    "value": [10, 20, 30, 40, 50]
})

    time  value
0    0hr     10
1  128hr     20
2   72hr     30
3   48hr     40
4   96hr     50

from natsort import natsort_keygen

df.sort_values(
    by="time",
    key=natsort_keygen()
)

    time  value
0    0hr     10
3   48hr     40
2   72hr     30
4   96hr     50
1  128hr     20