Get pandas series labels of true values without storing the series in a temp variable

Question

Input:

pd.Series([True, False, True, False], index=['a', 'b', 'c', 'd'])

Desired result:

['a', 'c']  # can be in pd.Series or np.array format, doesn't matter.

But I want a solution which does not store the series in a temporary variable, e.g.

s = pd.Series([True, False, True, False], index=['a', 'b', 'c', 'd'])
s[s == True].index

stores the series in s, and uses it twice.

The solution I'm looking for is like

np.where(s)

But returns the label of the true values, rather than their integer indices.

This question is similar to this one, but more specific.

score 4 · Accepted Answer · answered Jun 27 '20 at 19:46

4

I always pass to loc

pd.Series([True, False, True, False], index=['a', 'b', 'c', 'd']).loc[lambda x : x].index
Index(['a', 'c'], dtype='object')

answered Jun 27 '20 at 19:46

BENY

Aha! Great answer. So there's no builtin for this in pandas? – Alex Lenail Jun 27 '20 at 20:24
@AlexLenail I do not think so ~ – BENY Jun 27 '20 at 20:26

1 Answers1