Unexpected behaviour in pandas dataframe .loc[] about add-trailing-comma HOT 3 CLOSED

df.loc[s, ]

and

df.loc[s]

are different code

the former is a 1-element tuple -- perhaps you meant to write df.loc[(s,)] ?

from add-trailing-comma.

Thanks - I thought it might be something like that, but I couldn't get it straight in my head. I guess it's a quirk/design decision of pandas that df.loc[s] and df.loc[s,]/df.loc[(s,)] should mean the same thing.

I suppose that the correct way to fix the conflict between flake8 (which prefers foo[1, ]) and this tool (which prefers foo[1,]) is to manually wrap the term in parentheses, i.e. foo[(1,)] - this seems to satisfy both tools. (Or preferably remove the comma entirely, i.e. foo[1], if that doesn't change the behaviour.)

from add-trailing-comma.

pandas may treat them the same, but they are not the same:

>>> class C:
...     def __getitem__(self, x):
...         print(x)
... 
>>> c = C()
>>> c[1,]
(1,)
>>> c[1]
1

from add-trailing-comma.