fetch_openml difference between pandas and liac-arff parser #23381
Description
Seen in a scipy-dev build.
There are additional quotes in the pandas parser case.
cc @glemaitre
import numpy as np
from sklearn.datasets import fetch_openml
mice_pandas = fetch_openml(name='miceprotein', version=4, parser="pandas")
mice_liac_arff = fetch_openml(name='miceprotein', version=4, parser="liac-arff")
assert np.testing.assert_array_equal(mice_pandas.target, mice_liac_arff.target)Output:
AssertionError:
Arrays are not equal
Mismatched elements: 1080 / 1080 (100%)
x: array(["'c-CS-m'", "'c-CS-m'", "'c-CS-m'", ..., "'t-SC-s'", "'t-SC-s'",
"'t-SC-s'"], dtype=object)
y: array(['c-CS-m', 'c-CS-m', 'c-CS-m', ..., 't-SC-s', 't-SC-s', 't-SC-s'],
dtype=object)