Using user defined structure as precision to fread

http://www.mathworks.com/help/techdoc/ref/fread.html

Above lists all the possible precision value. They are the MATLAB defined data types.

Say the binary file i am reading follows the following repetitive basic structure: floatValue, 4byte, float type doubleValue, 8byte, double type

I define a structure: myStructure = struct('floatValue', single(213.5883), 'doubleValue', 57605.061382);

Now i want to be able to read my binary file as follows: fid = fopen('myBinaryFile.ext'); A = fread(fid,1,*'myStructure'*,'s'); %Read one element data

Finally i can access the data as: A.floatValue or A(1).floatValue A.doubleValue or A(1).doubleValue

Concern is reading in the custom structure format. Above gives a invalid precision error. What way can we read our cutsom structures?

Answer 1

[Walter Roberson]

To read repetitive binary structures, memmapfile() is the appropriate routine.

[Akash]

Thanks that worked! :)

[gujax]

I would like to follow up on this...

I am not sure if Matlab's fread is derived from C++'s fread?

C++'s fread or even istream can read an array of structured data (I just learned it from chatGPT). So why Matlab's fread is limited to a fixed type read?

Answer 2

[Kaustubha Govind]

There is no precision format to directly read into a structure. You need to read the file into arrays/matrices and then later compile them into a structure. Perhaps TEXTSCAN will be more appropriate for youm because it allows you to use a format string that is composed of several different types - the values are read into a cell array, which can store disparate types.

[Akash]

Thanks for your reply. Memmapfile solved my problem.