Appending np.nan to bool ndarray changes its dtype

[kaelzhang]

Tell us about your environment

• Platform (Windows / MacOs / Linux) and version: macOS
• What kind of virtual env do you use ? (virtualenv / anaconda / None): conda

Run the following command and paste the output

python -V
pip list | grep pandas

Python 3.8.1
pandas             1.0.3
stock-pandas       0.23.1

code

stock['kdj.j / kdj.d'] should be of bool dtype, but appending np.nan to the series will change its dtype to np.float64

Actually, it is not a bug of stock-pandas, but the behavior of numpy