C Syntax

#include "matrix.h"
mxArray *mxCreateNumericArray(mwSize ndim, const mwSize *dims, 
         mxClassID classid, mxComplexity ComplexFlag);

Fortran Syntax

mwPointer mxCreateNumericArray(ndim, dims, classid, 
  ComplexFlag)
mwSize ndim
mwSize dims(ndim)
integer*4 classid, ComplexFlag

Arguments

ndim

Number of dimensions. If you specify a value for ndim that is less than 2, mxCreateNumericArray automatically sets the number of dimensions to 2.

dims

Dimensions array. Each element in the dimensions array contains the size of the array in that dimension. For example, in C, setting dims[0] to 5 and dims[1] to 7 establishes a 5-by-7 mxArray. In Fortran, setting dims(1) to 5 and dims(2) to 7 establishes a 5-by-7 mxArray. In most cases, there are ndim elements in the dims array.

classid

Identifier for the class of the array, which determines the way the numerical data is represented in memory. For example, specifying mxINT16_CLASS in C causes each piece of numerical data in the mxArray to be represented as a 16-bit signed integer. In Fortran, use the function mxClassIDFromClassName to derive the classid value from a MATLAB^® class name. See the Description section for more information.

ComplexFlag

If the mxArray you are creating is to contain imaginary data, set ComplexFlag to mxCOMPLEX in C (1 in Fortran). Otherwise, set ComplexFlag to mxREAL in C (0 in Fortran).

Returns

Pointer to the created mxArray, if successful. If unsuccessful in a standalone (non-MEX file) application, returns NULL in C (0 in Fortran). If unsuccessful in a MEX file, the MEX file terminates and returns control to the MATLAB prompt. The function is unsuccessful when there is not enough free heap space to create the mxArray.

Description

Call mxCreateNumericArray to create an N-dimensional mxArray in which all data elements have the numeric data type specified by classid. After creating the mxArray, mxCreateNumericArray initializes all its real data elements to 0. If ComplexFlag equals mxCOMPLEX in C (1 in Fortran), mxCreateNumericArray also initializes all its imaginary data elements to 0. mxCreateNumericArray differs from mxCreateDoubleMatrix as follows:

• All data elements in mxCreateDoubleMatrix are double-precision, floating-point numbers. The data elements in mxCreateNumericArray can be any numerical type, including different integer precisions.

• mxCreateDoubleMatrix can create two-dimensional arrays only; mxCreateNumericArray can create arrays of two or more dimensions.

mxCreateNumericArray allocates dynamic memory to store the created mxArray. When you finish with the created mxArray, call mxDestroyArray to deallocate its memory.

MATLAB automatically removes any trailing singleton dimensions specified in the dims argument. For example, if ndim equals 5 and dims equals [4 1 7 1 1], the resulting array has the dimensions 4-by-1-by-7.

The following table shows the C classid values and the Fortran data types that are equivalent to MATLAB classes.

Examples

See the following examples in matlabroot/extern/examples/refbook.

• phonebook.cphonebook.c

• doubleelement.cdoubleelement.c

• matrixDivide.cmatrixDivide.c

• matsqint8.Fmatsqint8.F

See the following examples in matlabroot/extern/examples/mx.

• mxisfinite.cmxisfinite.c

See Also

mxClassId, mxClassIdFromClassName, mxComplexity, mxDestroyArray, mxCreateUninitNumericArray, mxCreateNumericMatrix