Feature request: Allow the subset argument of dropna and drop_duplicates to take a single label

[tsoernes]

This works:

df.drop(columns='My Col')

But neither of these do:

df.dropna(subset='My Col')
df.drop_duplicates(subset='My Col')
df.duplicated(subset='My Col')

all of which instead should be called with subset=['My Col'].
This is a feature request for the three latter functions to allow the subset argument to take a single label.

[WillAyd]

I see your argument but I think it’s actually problematic that the first one works. Not sure it’s worth expanding this behavior and might even lean towards deprecating passing a single string to “columns”

[MarcoGorelli]

might even lean towards deprecating passing a single string to “columns”

For what it's worth, I would love to see that happen