Convert an array stored as a string, to a proper numpy array

Question

For long and tedious reasons, I have lots of arrays that are stored as strings:

tmp = '[[1.0, 3.0, 0.4]\n [3.0, 4.0, -1.0]\n [3.0, 4.0, 0.1]\n [3.0, 4.0, 0.2]]'

Now I obviously do not want my arrays as long strings, I want them as proper numpy arrays so I can use them. Consequently, what is a good way to convert the above to:

tmp_np = np.array([[1.0, 3.0, 0.4]
 [3.0, 4.0, -1.0]
 [3.0, 4.0, 0.1]
 [3.0, 4.0, 0.2]])

such that I can do simple things like tmp_np.shape = (4,3) or simple indexing tmp_np[0,:] = [1.0, 3.0, 0.4] etc.

Thanks

• Kasravnd: Have you seen stackoverflow.com/questions/20200353/… and a lot of other documentations regard this problem in Internet?

Answer 1

You can use ast.literal_eval, if you replace your \n characters with ,:

temp_np = np.array(ast.literal_eval(tmp.replace('\n', ',')))

Returns:

>>> tmp_np
array([[ 1. , 3. , 0.4],
 [ 3. , 4. , -1. ],
 [ 3. , 4. , 0.1],
 [ 3. , 4. , 0.2]])