Numpy issubdtype gives “TypeError: data type not understood”

Question

I'm running python 2.7 and numpy 1.15.

I get:

>>> import numpy as np
>>> np.issubdtype(4, float)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/brianp/work/cyan/venv/lib/python2.7/site-packages/numpy/core/numerictypes.py", line 714, in issubdtype
    arg1 = dtype(arg1).type
TypeError: data type not understood

Did something change so that it used to work with values, but now it only works with types?

As per an answer below:

>>> np.issubdtype(4, np.float)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/brianp/work/cyan/venv/lib/python2.7/site-packages/numpy/core/numerictypes.py", line 714, in issubdtype
    arg1 = dtype(arg1).type
TypeError: data type not understood

I should add that

>>> np.issubdtype(type(4), np.float)
False

works... but the code USED to work without type() ...

Answer 1

Use the numpy types first. Then you want to compare dtypes, not actual values.

>>> np.issubdtype(np.float, np.float)
True

You cannot compare variables, only types.