cut: documentation wrongly claims that labels=False (always) returns an ndarray of ints

[juliangilbey]

The documentation for pandas.cut says that the output return type depends on the value of the labels parameter, and when this is False:

pandas/pandas/core/reshape/tile.py

Line 130 in 93b5135

* False : returns an ndarray of integers.

But if the input to cut is a Series, then the output is as well, either with dtype=int64 if there are no NaNs in the result, or float64 if there are; an example of this behaviour appears in the documentation for the function.

[juliangilbey]

I forgot to give an example. Here is one giving a Series output with int64; there is an example with float64 in the pandas.cut documentation:

s = pd.Series([0.7, 0.1, -0.4, 0.3, 1.2, -0.8])
bins = [-10, -0.5, 0, 0.5, 10]
pd.cut(s, bins, labels=False)

[Mike-gag]

take