How can I write a numpy.array to a file and retain all of the digits?

Question

Suppose I have a numpy array with some high precision floats obtained from scipy.optimize.minimize, such as: arr = np.array([9.2387213981273981, 0.3219837123801298]). I would like to write this array with this level of precision to a file in a way like the following:

filename = 'parameters.py'
f = open(filename, 'w')
f.write('from numpy import *' + '\n')
f.write('arr = ' + repr(arr))
f.close()

However, this doesn't work, since repr(arr) returns 'array([9.2387214 , 0.32198371])', and there is a high level of precision loss here. One potential way to fix this is to convert arr to a list and write the list to the file instead:

L = list(arr)
filename = 'parameters.py'
f = open(filename, 'w')
f.write('L = ' + repr(L))
f.close()

Here, repr(L) returns '[9.238721398127398, 0.3219837123801298]', and so there is no precision loss. My question is, how can I write the array to a file without converting it to a list, so that the there is no precision loss?

Text isn't a good format for this purpose.

hpaulj
– hpaulj

2021-07-04 20:56:37 +00:00
Commented Jul 4, 2021 at 20:56 — hpaulj
– hpaulj, Commented Jul 4, 2021 at 20:56

Mustafa Aydın · Accepted Answer · 2021-07-04 19:44:57Z

1

You can use np.printoptions context manager and set the precision to, say, 15:

with np.printoptions(precision=15), open(filename, "w") as fh:
    fh.write("from numpy import *\n")
    fh.write("arr = " + repr(arr))

after which the file looks like:

from numpy import *
arr = array([9.238721398127398, 0.32198371238013 ])

answered Jul 4, 2021 at 19:44

Mustafa Aydın

18.4k4 gold badges21 silver badges43 bronze badges

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

1 Comment

Mustafa Aydın Over a year ago

(also noting the np.save and np.load functions.)

Collectives™ on Stack Overflow

How can I write a numpy.array to a file and retain all of the digits?

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related