NumPy ndarray.all () vs np.all (ndarray) vs all (ndarray)

Question

What is the difference between the three “all” methods in Python / NumPy? What is the reason for the difference in performance? Is it true that ndarray.all () is always the fastest of the three?

Here is the time test that I ran:

In [59]: a = np.full(100000, True, dtype=bool)

In [60]: timeit a.all()
The slowest run took 5.40 times longer than the fastest. This could mean that an intermediate result is being cached.
100000 loops, best of 3: 5.24 µs per loop

In [61]: timeit all(a)
1000 loops, best of 3: 1.34 ms per loop

In [62]: timeit np.all(a)
The slowest run took 5.54 times longer than the fastest. This could mean that an intermediate result is being cached.
100000 loops, best of 3: 6.41 µs per loop

Answer 1

The difference between np.all(a)and a.all()is simple:

• If it ais numpy.array, then np.all()it will simply call a.all().
• a numpy.array, np.all() numpy.array, a.all(). a.all() , a numpy.array , , all.

np.all all .

• all ( list, set s, generators,...). np.all numpy.array ( , numpy, .. list tuple s).
• np.all array , != 0. all bool , .
• python , python. np.all .

, a. python, all . , np.all a.all() ( , , object , , ).

Answer 2

• np.all - , , , , ndarray.all, .

• all - python, . https://docs.python.org/2/library/functions.html#all.

• ndarray.all - numpy.ndarray, .

Answer 3

I suspect that numpy functions do more to evaluate the element of the array as logical, probably in some generic numerical order, while the built-in all()does nothing, since the elements are already logical.

I wonder how different results will be with integers float.