numpy "TypeError: ufunc 'bitwise_and' not supported for the input types" and the inputs could not be safely coerced to any supported types

Question

my input dataframe:

    F1  F2  F3  F4  F5  F6  F7  F8  F9  F10 F11 F12 F13 F14 F15 F16
 F1 1.0 1.0 1.0 1.0 1.0 1.0 0.0 1.0 0.0 0.0 0.0 1.0 1.0 0.0 1.0 1
 F2 1.0 1.0 1.0 0.0 1.0 0.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 0.0 0.0 1
 F3 0.0 1.0 1.0 1.0 0.0 1.0 1.0 0.0 1.0 0.0 0.0 1.0 1.0 1.0 1.0 1
 F4 0.0 0.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 0.0 1.0 1.0 1.0 1.0 0
 F5 0.0 0.0 0.0 1.0 1.0 1.0 1.0 1.0 0.0 1.0 1.0 0.0 0.0 1.0 1.0 1
 F6 0.0 0.0 0.0 0.0 0.0 1.0 0.0 1.0 1.0 1.0 1.0 1.0 1.0 0.0 1.0 1
 F7 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 0.0 1
 F8 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 0
 F9 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1
F10 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 1.0 1.0 1.0 1.0 1
F11 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 1.0 1.0 1.0 1.0 1
F12 0.0 0.0 0.0 1.0 1.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0 0.0 1.0 1.0 1
F13 1.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 1.0 1
F14 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 1.0 0.0 0.0 1.0 0.0 1.0 0.0 1
F15 0.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 1.0 0.0 1.0 0
F16 0.0 0.0 1.0 0.0 0.0 1.0 1.0 1.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 1

I want to apply this code:

m = df2.to_numpy()

mask = m * m[:, np.newaxis, :].transpose((0,2,1))

bol = mask.any(axis=1)

final = m | bol

But at the final step I get this error:

TypeError: ufunc 'bitwise_or' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule ''safe''

Please tell me where is the mistake?

Answer 1

Convert values of array m to boolean for mapping 1/0 to True/False :

final = m.astype(bool) | bol

---

print (final)
[[ True  True  True  True  True  True  True  True  True  True  True  True
   True  True  True  True]
 [ True  True  True  True  True  True  True  True  True  True  True  True
   True  True  True  True]
 [ True  True  True  True  True  True  True  True  True  True  True  True
   True  True  True  True]
 [ True  True  True  True  True  True  True  True  True  True  True  True
   True  True  True  True]
 [False  True  True  True  True  True  True  True  True  True  True  True
   True  True  True  True]
 [ True  True  True  True  True  True  True  True  True  True  True  True
   True  True  True  True]
 [ True  True  True  True  True  True  True  True  True  True  True  True
   True  True  True  True]
 [ True  True  True  True  True  True  True  True  True  True  True  True
   True  True  True  True]
 [ True  True  True  True  True  True  True  True  True  True  True  True
   True  True  True  True]
 [ True  True  True  True  True  True  True  True  True  True  True  True
   True  True  True  True]
 [ True  True  True  True  True  True  True  True  True  True  True  True
   True  True  True  True]
 [False  True  True  True  True  True  True  True  True  True  True  True
   True  True  True  True]
 [ True  True  True  True  True  True  True  True  True  True  True  True
   True False  True  True]
 [ True False  True  True  True  True  True  True  True  True  True  True
   True  True  True  True]
 [ True  True  True  True  True  True  True  True  True  True  True  True
   True  True  True  True]
 [False  True  True  True False  True  True  True  True  True  True  True
   True  True  True  True]]

Last cast to integers:

final = (m.astype(bool) | bol).astype(int)

Answer 2

m is an array of floats, you cannot directly perform a A | B A | B operation if not both arrays are of boolean type.

Assuming you want to consider 0 / 1 as False / True , use:

m = df2.to_numpy().astype(bool)

or

final = m.astype(bool)|bol

If you only need final and not the intermediates, it is more efficient to convert to boolean early.

Output as integer:

final = (m.astype(bool)|bol).astype(int)

or, if early conversion of m to boolean:

# m = df2.to_numpy().astype(bool)
# ...
final = (m|bol).astype(int)