Pass Data Between MATLAB Production Server and Python

R2026b

Pass Data from MATLAB to Python

When you pass data from MATLAB^® to Python^®, either directly or as output from a MATLAB function in Python, MATLAB converts the data into the equivalent Python data types.

MATLAB Output Argument Type (scalar unless otherwise noted)	Resulting Python Data Type
Numeric array	matlab numeric array object See MATLAB Arrays as Python Variables (MATLAB).
double single	float
Complex (any numeric type)	complex
int8, int16, int32, int64 uint8, uint16, uint32, uint64	int
NaN	float('nan')
Inf	float('inf')
logical	bool
char array (1-by-N or N-by-1)	str
string	str
string array (1-by-N or N-by-1)	list of str
struct	dict
cell array (1-by-N or N-by-1)	list
dictionary	matlab.dictionary
datetime	datetime.datetime
duration	datetime.timedelta
datetime array	numpy.datetime64 array If the NumPy module is not available in the Python environment, MATLAB converts the datetime array to matlab.object.
duration array	numpy.timedelta array If the NumPy module is not available in the Python environment, MATLAB converts the duration array to matlab.object.
table timetable	pandas.DataFrame Only one level of nesting in tables is supported. If the Python Pandas module is not available in the Python environment, MATLAB converts the table or timetable to a matlab.object.
MATLAB handle object (such as the containers.Map type)	matlab.object MATLAB returns a reference to a matlab.object, not the object itself. You cannot pass a matlab.object between MATLAB sessions.
MATLAB value object (such as the categorical type)	matlab.object You can pass a value object to a MATLAB function, but you cannot create or modify it.

Unsupported MATLAB Data Types

Along with the exceptions noted in the table, MATLAB does not convert these MATLAB data types to Python data types:

• char array (M-by-N)

• string array (M-by-N)

• cell array (M-by-N)

• Sparse array

• struct array

• Objects not produced by MATLAB (such as Java^® objects)

• Function handle

Pass Data from Python to MATLAB

When you pass data from Python to MATLAB, either directly or as input arguments to a MATLAB function in Python, MATLAB Production Server™ converts the data into the equivalent MATLAB data types.

Python Data Type	Resulting MATLAB Data Type (scalar unless otherwise noted)
matlab numeric array object	Numeric array See matlab Python Module.
Python objects that support the buffer protocol, such as memoryview, array.array, or NumPy ndarray	Numeric array
float	double
complex	Complex double
int	int64
float('nan') numpy.nan	NaN
float('inf') numpy.inf	Inf
bool	logical
str	char
bytearray	uint8 array
bytes	uint8 array
list	cell array
set	cell array
tuple	cell array
dict	struct dict keys must meet certain specifications. See Use MATLAB Dictionaries in Python (MATLAB).
matlab.dictionary	dictionary matlab.dictionary keys and values must meet certain specifications. MATLAB converts the data in a matlab.dictionary object using the mappings in this table, with the exception that str data is converted to string. See Use MATLAB Dictionaries in Python (MATLAB).
datetime.datetime numpy.datetime64 scalar	datetime
datetime.timedelta numpy.timedelta64 scalar	duration
numpy.datetime64 array	datetime array
numpy.timedelta64 array	duration array
pandas.DataFrame	table Only one level of nesting in tables is supported.
pandas.DataFrame with datetime column	timetable Only one level of nesting in timetables is supported.
Python objects other than dict that support the mapping protocol (such as collections.OrderedDict)	dictionary
matlab.object	When you pass a MATLAB object that does not have an explicit conversion rule to Python, MATLAB passes it as a matlab.object. You can return the matlab.object back to MATLAB, in which case it is treated like the original MATLAB object.

Unsupported Python Data Types

Along with the exceptions noted in the table, MATLAB does not convert these Python data types to MATLAB data types:

• Python class (module.type) objects

• None object

• enum type