HDF5 - reading part of dataset

I'd like to read only part of the dataset of HDF5 file using h5read function in Matlab . Particularly in 1 of the dimensions I want to read only every 4th value.

I can read the whole dataset as:

tempp = hdf5read(nameCur,namedataset);

its size: 160x100x200x10.

Next, when I want to read the whole dataset using "start" and "count" attributes, it works fine:

tempp = h5read(nameCur,namedataset,[1 1 1 1],[160 100 200 10]);

Then I want to read all dataset using "stride" attribute too:

tempp = h5read(nameCur,namedataset,[1 1 1 1],[160 100 200 10],[1 1 1 1]);

works nice!

But when I want finally to read every 4th value from second dimension

tempp = h5read(nameCur,namedataset,[1 1 1 1],[160 100 200 10],[1 4 1 1]);

it gives me error:

Error using h5readc

The index arguments exceed the size of the dataset.

I don't understand what is the reason for that. Thanks

[Pavel Inchin]

Just in case someone stucked with this, the solution is simply:

tempp = h5read(nameCur,namedataset,[1 1 1 1],[160 25 200 10],[1 4 1 1]);

Answer 1

[Sunson29]

Dear friend, I hope I could answer your question. But I guess I am more "beginer " than you.

May I ask a very quick question ?

I am trying to open just one portion of HDF5 file, because it's way too big (20G). In that files, there are 3 dataset, X,Y,Z, X is the 20G, others are small, can be open directly.

I can run the code "dset = hdf5read('GOLD_XYZ_OSC.0001_1024.hdf5','/Z');";

but why I cannot run the code like you do dset = hdf5read('GOLD_XYZ_OSC.0001_1024.hdf5','/Z', [1 1], [1 3]);

Error: Expected a string scalar or character vector for the parameter name.

Z is a big row, one row, many column. I tried to put just few elements (i=1:3) in dset. Thank you!!!!

[Pavel Inchin]

what is a size of "Z" you'd like to read?

Answer 2

[Sunson29]

my friend, you reply so fast!

Could you go to this link plz? It's easier to see.

https://www.mathworks.com/matlabcentral/answers/456544-how-to-read-only-one-potion-of-hdf5-files-in-matlab

I got very very similiar index issue, like you had before. Thank you!!!!