Main Content

MATLAB to Python Data Type Mapping

When calling a Python® function, MATLAB® converts MATLAB data into types that best represent the data to the Python language.

Pass Scalar Values to Python

MATLAB Input Argument Type —
Scalar Values Only

Resulting Python py. Type

Examples

double (real)
single (real)

float

Use Python Numeric Variables in MATLAB

double (complex)
single (complex)

complex

z = complex(1,2);
py.cmath.polar(z)
ans = 
  Python tuple with no properties.

    (2.23606797749979, 1.1071487177940904)

int8
uint8
int16
uint16
int32

int

 

uint32
int64
uint64

int
long (version 2.7 only)

 

NaN

float("nan")

 

Inf

float("inf")

 

string scalar
char vector

str

Use Python str Variables in MATLAB

<missing> value in string

None

py.list({string(missing),'Value'})
ans = 
  Python list with no properties.

    [None, 'Value']

logical

bool

 

Structure

dict

Use Python dict Variables in MATLAB

Python object — py.type

type

 

function handle @py.module.function, to Python functions only

module.function

Pass Python Function to Python map Function

Pass Vectors to Python

MATLAB Input Argument Type —
1-by-N Vector

Resulting Python Type

double (real)

array.array('d')

single (real)

array.array('f')

int8 (real)

array.array('b')

uint8 (real)

array.array('B')

int16 (real)

array.array('h')

uint16 (real)

array.array('H')

int32 (real)

array.array('i')

uint32 (real)

array.array('I')

int64 (real) - Not supported for Python 2.7 on Windows®

array.array('q')

uint64 (real) - Not supported for Python 2.7 on Windows

array.array('Q')

double (complex)
single (complex)
int8 (complex)
uint8 (complex)
int16 (complex)
uint16 (complex)
int32 (complex)
uint32 (complex)

memoryview

logical

memoryview

char vector
string scalar

str

char array containing values greater than 127 (version 2.7 only)

unicode

cell vector

tuple

Pass Matrices and Multidimensional Arrays to Python

The Python language provides a protocol for accessing memory buffers like the data stored in a MATLAB array. MATLAB implements this Python buffer protocol for MATLAB arrays so that you can read MATLAB arrays directly from Python code, running in the same process as MATLAB, without copying data.

Many Python functions directly use the MATLAB array from Python without converting it to a native Python type. Some functions might require a specific type, such as numpy.ndarray, or might modify data in the array. These functions might accept the MATLAB array and copy the data into the required type. Other functions might display an error if you do not pass the required type. To pass data to these functions, first create the required Python type from the MATLAB data, then pass it to the Python function. For example, to create array p to pass to a Python function that requires data of type numpy.array, type:

p = py.numpy.array(magic(3))
p = 

  Python ndarray:

     8     1     6
     3     5     7
     4     9     2

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

    Use double function to convert to a MATLAB array.

MATLAB sparse arrays are not supported in Python. See Unsupported MATLAB Types.

Troubleshooting Argument Errors

If a Python function expects a specific Python multidimensional array type such as numpy.ndarray, then MATLAB displays a message with tips about how to proceed. If the problem might be due to passing a matrix or a multidimensional array as an argument, then do the following.

  1. Check the documentation for the Python function and find out the expected type for the argument.

  2. Create a Python object of that type in MATLAB and pass that to the Python function.

For example, suppose that the following code returns an error.

a = [1 2; 3 4];
py.pyfunc(a)

If the documentation of pyfunc specifies that the expected type is numpy.ndarray, then try this conversion:

py.pyfunc(numpy.ndarray(a))

If the error persists, then determine the root cause by checking for additional information in the Python exception.

Automatically Convert Python Types to MATLAB Types

MATLAB automatically converts these data types returned from Python into MATLAB types. To convert other types, see Explicitly Convert Python Types to MATLAB Types.

Python Return Type, as Displayed in Python

Resulting MATLAB Type — Scalar

float

double

complex

Complex double

int (version 2.7 only).

For Python versions 3.x int, you must convert explicitly. See Explicitly Convert Python Types to MATLAB Types.

int64

bool

logical

All other Python types — type

Python object — py.type

Explicitly Convert Python Types to MATLAB Types

If the output of a Python function implements the Python buffer protocol, such as numpy.ndarray, and it is numeric or logical, then MATLAB displays:

  • The actual Python type

  • The underlying data

  • The corresponding MATLAB conversion function. Use this function to fully convert the Python object to a MATLAB array.

Use these MATLAB functions to convert Python data types to MATLAB types.

Python Return Type or Protocol, as Displayed in MATLAB

MATLAB Conversion Function

Examples

py.str (version 3.x)

string
char

Use Python str Variables in MATLAB

py.str (version 2.7)

string
char
uint8

 

py.unicode

string
char

 

Object with __str__ method

char

py.help('datetime.date.__str__')
Help on wrapper_descriptor in datetime.date:

datetime.date.__str__ = __str__(self, /)
    Return str(self).
d = py.datetime.date(...
    int32(2020),int32(3),int32(4));
char(d)
ans = '2020-3-04'

py.bytes

uint8

 

py.int

double
or
int64

 

py.long

double
or
int64

 

py.array.array
py.numpy.ndarray
py.memoryview

You can convert py.array.array of any format and py.memoryview objects to the MATLAB type you want.

numeric
double
single
int8
uint8
int16
uint16
int32
uint32
int64
uint64

Use Python Numeric Variables in MATLAB, for example, Use Python Integer array Types in MATLAB.

Sequence protocol; for example, py.list and py.tuple

cell

Use Python list Variables in MATLAB
Use Python tuple Variables in MATLAB

Mapping protocol; for example, py.dict

struct

Use Python dict Variables in MATLAB

For example, a Python function returns this array p:

p = 

  Python ndarray:

     8     1     6
     3     5     7
     4     9     2

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

    Use double function to convert to a MATLAB array.

You can convert it to a MATLAB matrix P by typing:

P = double(p)
P = 3×3    
     8     1     6
     3     5     7
     4     9     2

If you need specific information about the Python properties of p, type:

details(p)
    py.numpy.ndarray handle with properties:

           T: [1×1 py.numpy.ndarray]
        base: [1×1 py.NoneType]
      ctypes: [1×1 py.numpy.core._internal._ctypes]
        data: [1×3 py.memoryview]
       dtype: [1×1 py.numpy.dtype[float64]]
       flags: [1×1 py.numpy.flagsobj]
        flat: [1×1 py.numpy.flatiter]
        imag: [1×1 py.numpy.ndarray]
    itemsize: [1×1 py.int]
      nbytes: [1×1 py.int]
        ndim: [1×1 py.int]
        real: [1×1 py.numpy.ndarray]
       shape: [1×2 py.tuple]
        size: [1×1 py.int]
     strides: [1×2 py.tuple]

  Methods, Events, Superclasses

If the Python module provides content in its __doc__ attribute, then MATLAB links to that information.

Unsupported MATLAB Types

These MATLAB types are not supported in Python.

  • Multidimensional char or cell arrays

  • Structure arrays

  • Sparse arrays

  • categorical,
    table,
    containers.Map,
    datetime types

  • MATLAB objects

  • meta.class (py.class)

Related Topics