How to pretty-print a tuple of NumPy scalars?

Question

This NumPy 2.2.6 script:

import numpy as np 

a = np.arange(24).reshape(2, 2, 2, 3)
idx = np.unravel_index(np.argmax(a, axis=None), a.shape)
print(idx)

will print:

(np.int64(1), np.int64(1), np.int64(1), np.int64(2))

which is hard to read. What is a simple way to print idx as (1, 1, 1, 2)? I'm looking for something better than print(np.array(idx)) workaround.

tuple(i.item() for i in idx); you're printing a literal tuple (try type(idx)) — Kenzo Staelens
– Kenzo Staelens, Commented Sep 1 at 6:47
@Daweo There has been a change in the scalar representation as of Numpy 2.0.0 (see release notes): As per NEP 51, the scalar representation has been updated to include the type information to avoid confusion with Python scalars. — simon
– simon, Commented Sep 1 at 7:17

mozway · Accepted Answer · 2025-09-01 07:51:02Z

4

This is the new way to represent numpy scalars since version 2.

You can revert to the old printing style with the legacy parameter of numpy.set_printoptions:

np.set_printoptions(legacy='1.25')

print(idx)
# (1, 1, 1, 2)

And to switch back to the new style: np.set_printoptions(legacy=False).

Alternatively, explicitly convert to python integers or strings:

print(tuple(int(i) for i in idx))  # or: print(tuple(map(int, idx)))
# (1, 1, 1, 2)

answered Sep 1 at 7:51

mozway

267k13 gold badges56 silver badges106 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Kelly Bundy · Accepted Answer · 2025-09-02 20:20:15Z

3

Since you said printing 1 1 1 2 would be ok, too:

print(*idx)

That prints each element's str instead of its repr.

answered Sep 2 at 20:20

Kelly Bundy

28k7 gold badges34 silver badges76 bronze badges

Collectives™ on Stack Overflow

How to pretty-print a tuple of NumPy scalars?

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related