API: Series.__setitem__ with Series value

[jbrockmendel]

orig = pd.Series(range(10))

key = pd.Series([6, 7])

ser = orig.copy()
ser[key._values] = key * 2
>>> ser
0     0
1     1
2     2
3     3
4     4
5     5
6    12
7    14
8     8
9     9
dtype: int64

ser = orig.copy()
ser.loc[key._values] = key * 2
>>> ser
0    0.0
1    1.0
2    2.0
3    3.0
4    4.0
5    5.0
6    NaN
7    NaN
8    8.0
9    9.0
dtype: float64

In .loc we do a .reindex on the value Series, but not in Series.__setitem__ (which goes through Series._set_values to mgr.setitem) using .iloc gets the __setitem__ behavior.

Maybe I just need some coffee, but it isn't obvious to me what the expected behavior is here cc @phofl

[phofl]

Loc is correct I think, they should align by your index (at least in the DataFrame case when setting values). If there is no reason to treat Series differently, then I'd keep the loc behavior

[jbrockmendel]

So you’d deprecate the series set item behavior or keep all three as-is?

[phofl]

If there is no reason that I am not aware of to keep the series set item behavior, I’d deprecate it, yes

[jbrockmendel]

Possible duplicate of #37516?

[topper-123]

I agree doing it the same as in loc.

A slightly related issue is that neither the doc string, see section "Setting values" nor the docs, see below the paragraph "Note that setting works as well:" explain that setting using series uses reindexing, it can catch some users off guard.