Initialization of Woodwork DataTable using pandas DataFrame and then numpy array causes different behavior from just initalization using numpy array.

[angela97lin]

If I initialize a Woodwork DataTable using a pandas DataFrame and then initialize another Woodwork DataTable using the numpy array underneath, it creates a Woodwork DataTable with category types. However, skipping the first step creates a Woodwork DataTable of ints/floats.

Printing out the underlying numpy array, it looks like the numpy array gets converted s.t. it has a dtype=object, causing Woodwork to convert to category types, but this behavior is confusing / inconsistent.

Code Sample, a copy-pastable example to reproduce your bug.

import woodwork as ww
import pandas as pd
X = pd.DataFrame([[1, 0]])
ww.DataTable(X)
ww.DataTable(X.to_numpy())

vs

import woodwork as ww
import pandas as pd
X = pd.DataFrame([[1, 0]])
# ww.DataTable(X)
ww.DataTable(X.to_numpy())

[tyler3991]

Let's revisit this after the accessor change.

[tyler3991]

When we address this one, also keep an eye on #456

[gsheni]

@angela97lin Is this still a problem or can we close?