One Hot Encoder errors out if there are missing values even if there are no categorical features

[freddyaboulton]

Repro

import pandas as pd
import woodwork as ww
from evalml.pipelines.components import OneHotEncoder
import pytest


df = pd.DataFrame({"a": [1.2, 2.3, 4.5, 6.7],
                   "b": [True, False, True, True],
                   "c": [4.5, 8.3, None, 4.3]})
df.ww.init(logical_types={"a": "Double",
                          "b": "Boolean",
                          "c": "Double"})

with pytest.raises(ValueError, match="Input contains NaN"):
    OneHotEncoder().fit_transform(df)

I would expect this to be a no-op since there are no categorical features in the data.

[freddyaboulton]

This is the root cause of #2967