Create a datastore from the sample file, mapredout.mat, which is the output file of the mapreduce function.

ds = datastore('mapredout.mat');

Get the default number of partitions.

n = numpartitions(ds)

n = 1

By default, there is only one partition in ds because it contains only one small file.

Partition the datastore and return the datastore corresponding to the first part.

subds = partition(ds,n,1);

Read the data in subds.

while hasdata(subds)
    data = read(subds);
end

Get a number of partitions to parallelize datastore access over the current parallel pool. You must have Parallel Computing Toolbox installed.

ds = datastore('mapredout.mat');

n = numpartitions(ds, gcp);

parfor ii=1:n
    subds = partition(ds,n,ii);
    while hasdata(subds)
        data = read(subds);
    end
end