Use Case 10

Partial I/O: support to access a sub-set of a dataset

(miguel)

It is helpful to be able to read or write part of a dataset. One application would be to read part of a large dataset into memory or extending a dataset by appending new measurements to a chunked storage layout. This can be achieved in HDF5 using hyperslab selection elements.

The ability to select part of a dataset is also important for implementing parallel I/O operations by accessing the data with independent processes as described in Use Case 17.