Easy memory-mapped arrays for large datasets 1.0

Code to generate n-dimensional memory-mapped arrays for efficiently accessing large array datasetsMemory-mapped arrays reside within binary files, but are mapped to MATLAB's working memory. They are particularly useful when handling large datasets, since it is possible to perform operations on a subset of the array without loading the whole thing into working memory.This is an improved version of David Omer's previous submission, with added support for multiple n-dimensional arrays of varying class, as well as improved documentation.Example useage:-----------------% set up some random arraysarray1 = rand(100,200,'single');array2 = randi(255,300,400,'uint8'); % make an empty binary file holderdatain = {'array1',size(array1),class(array1);'array2',size(array2),class(array2)};makeEmptyMMBinFile('binfile.bin',datain);% map the arrays to the empty binary fileMMA = makeMMarray('binfile.bin',true);MMA.Data.array1 = myarray1;MMA.Data.array2 = myarray2;MMarray1 = MMA.Data.array1;MMarray2 = MMA.Data.array2;

Requirements:	No special requirements
Platforms:	Matlab
Keyword:	Added,  Class,  David,  Documentation,  Multiple,  Support,  Useage,  Varying,  Version
Users rating:	0/10