python - string representation of a numpy array with commas separating its elements

Question

I have a numpy array, for example:

points = np.array([[-468.927,  -11.299,   76.271, -536.723],
                   [-429.379, -694.915, -214.689,  745.763],
                   [   0.,       0.,       0.,       0.   ]])

if I print it or turn it into a string with str() I get:

print w_points
[[-468.927  -11.299   76.271 -536.723]
 [-429.379 -694.915 -214.689  745.763]
 [   0.       0.       0.       0.   ]]

I need to turn it into a string that prints with separating commas while keeping the 2D array structure, that is:

[[-468.927,  -11.299,   76.271, -536.723],
 [-429.379, -694.915, -214.689,  745.763],
 [   0.,       0.,       0.,       0.   ]]

Does anybody know an easy way of turning a numpy array to that form of string?

I know that .tolist() adds the commas but the result loses the 2D structure.

Answer 1

Try using repr

>>> import numpy as np
>>> points = np.array([[-468.927,  -11.299,   76.271, -536.723],
...                    [-429.379, -694.915, -214.689,  745.763],
...                    [   0.,       0.,       0.,       0.   ]])
>>> print(repr(points))
array([[-468.927,  -11.299,   76.271, -536.723],
       [-429.379, -694.915, -214.689,  745.763],
       [   0.   ,    0.   ,    0.   ,    0.   ]])

If you plan on using large numpy arrays, set np.set_printoptions(threshold=np.nan) first. Without it, the array representation will be truncated after about 1000 entries (by default).

>>> arr = np.arange(1001)
>>> print(repr(arr))
array([   0,    1,    2, ...,  998,  999, 1000])

Of course, if you have arrays that large, this starts to become less useful and you should probably analyze the data some way other than just looking at it and there are better ways of persisting a numpy array than saving it's repr to a file...