Error running mex file for C code gen: Subscript indices must either be real positive integers or logicals

I have a function that I coverted to C usingt the coder in double and single successfuly. However when I converted to fixed point using fi-embedded with datatype table, the buildIstumented successed in generating the mex version file, but when I attempted to run the mex version of the function to log data for the showInstrumentation tool , I receive error message above complaining that subscript must be either real positive integers or logicals. The input arguments to the function checks out as embedded fi format. But I cant resolve the issue after doing extensive research on the blogs and mathworks for Asking questions - I see that many complained about this error but the not adequate answers were provided. I use the manual method to covert matlab function into C rather than the app itself. The error is refered to a function that I built to find the maximum from an array of input values .

Answer 1

[Hari]

Hi Maya Jubran,

I understand that you are converting a MATLAB function to C code for both double and single precision using MATLAB Coder. However, while converting to fixed-point using the "fi" function and a datatype table, you have encountered an error on running the generated MEX file.

This error occurs there's a mismatch between the expected integer-type indices and the actual indices used in your custom maximum-finding function. MATLAB requires indices to be real positive integers or logicals, and this requirement is strict when dealing with fixed-point arrays.

To address the issue, ensure that all indexing operations within your function use integer-type variables. If you're using fixed-point numbers as indices, convert them to integers first. Here's an example of how you might modify the indexing:

% Assume 'array' is your input fixed-point array and 'index' is a fixed-point number
index = fi(5, 1, 16, 0); % Example fixed-point index
intIndex = int32(index); % Convert to integer for indexing
maxValue = array(intIndex); % Use integer index to access array elements

Also, ensure that any custom functions that you have written for operations like finding the maximum value are compatible with fixed-point arithmetic and indexing.

Refer to the documentation of Fixed-Point Designer for more information on working with fixed-point numbers in MATLAB.

https://www.mathworks.com/help/fixedpoint/index.html

For understanding how to convert fixed-point numbers to integers for indexing, refer to the "fi" function documentation https://www.mathworks.com/help/fixedpoint/ref/fi.html

Hope this helps!