Converting matlab.double to python compatible format

The documentation on this page states that data of type:

MATLAB Input Argument Type —

1-by-N Vector

double (real)

Gets converted to:

Resulting Python Type

array.array('d')

Consequently, I thought this python code would work:

import matlab.engine
import array
me = matlab.engine.start_matlab()
array.array('d', me.rand(1, 2))

However, this returns:

TypeError: must be real number, not matlab.double

What is the most pythonic way to convert a standard matlab double array into a python fundamental data type? (e.g. list or array.array)

Answer 1

[Selena Mastrodonato]

This syntax py.array.array('d', rand(1,2)) should work.

py.array.array('d', rand(1,2))

ans = 
  Python array:

    0.6970    0.3219

    Use details function to view the properties of the Python object.

    Use double function to convert to a MATLAB array.

Or if you want a Python list: py.list(rand(1,2))

py.list(rand(1,2))

ans = 
  Python list with values:

    [0.3941618631519793, 0.8464541453090321]

    Use string, double or cell function to convert to a MATLAB array.

[Lyle Collins]

Hi @Selena Mastrodonato, I've unintentionally confused you because my code has matlab syntax highlighting instead of python syntax highlighting. I've edited my question to make the problem clearer.

[Selena Mastrodonato]

Hi Lyle, thank you for providing more informations.

You could try this command: me.rand(1, 2)[0].toarray(), you'll obtain array('d', [0.8147236863931789, 0.9057919370756192]).

Instead, if you want a python list you could use numpy, so try this: numpy.array(me.rand(1, 2)[0].toarray()) and you'll obtain [0.81472369 0.90579194].

If you have a multidimensional matrix, try to iterate over it to format data as you wish.

[Lyle Collins]

Thanks @Selena Mastrodonato! Much appreciated!