HDF5 Overview

HDF5 Fortran Framework

Introduction

HDF5 (Hierarchical Data Format 5) is a versatile and high-performance data management framework developed by the HDF Group. It provides a flexible and efficient way to store and organize large amounts of data, making it ideal for scientific and engineering applications. The HDF5 Fortran Framework is a set of libraries and tools that allow Fortran programmers to work with HDF5 files.

History

HDF5 was first released in 1998 and has since become widely adopted in the scientific community. It was designed to address the limitations of earlier versions of the HDF format, providing better performance, scalability, and ease of use. The HDF Group continues to actively develop and maintain the HDF5 software, ensuring its compatibility with modern computing environments.

Features

The HDF5 Fortran Framework provides a rich set of features for working with HDF5 files:

Hierarchical Structure: HDF5 files are organized in a hierarchical structure, allowing users to create groups and datasets to organize their data effectively.
Efficient Data Compression: HDF5 supports various compression algorithms, such as GZIP and SZIP, to reduce the size of the stored data without sacrificing performance.
Parallel I/O: HDF5 can be used in parallel computing environments, enabling multiple processes to read and write data simultaneously for improved performance.
Chunking: HDF5 allows users to define the size of data chunks, which can be stored and accessed independently. This feature enables efficient access to subsets of large datasets.
Metadata: HDF5 supports the storage of metadata, allowing users to attach additional information to datasets and groups.
Data Types: HDF5 provides a wide range of data types, including integers, floating-point numbers, strings, and more. Custom data types can also be defined.
Attributes: HDF5 allows users to attach attributes to datasets and groups, providing a convenient way to store additional information.
Data Filtering: HDF5 supports data filtering operations, such as data shuffling and data compression, to improve the efficiency of data storage and retrieval.

Examples

Example 1: Creating an HDF5 File

Here's an example of how to create a new HDF5 file using the HDF5 Fortran API:

program create_hdf5_file
  use hdf5
  implicit none

  integer(hid_t) :: file_id, group_id, dataspace_id, dataset_id, status
  integer(hsize_t) :: dims(2)
  real :: data(10, 10)

  ! Create a new HDF5 file
  status = H5Fcreate("example.h5", H5F_ACC_TRUNC, file_id)

  ! Create a group within the file
  group_id = H5Gcreate(file_id, "/group", H5P_DEFAULT, H5P_DEFAULT, H5P_DEFAULT)

  ! Create a dataset within the group
  dims = [10, 10]
  dataspace_id = H5Screate_simple(2, dims, dims)
  dataset_id = H5Dcreate(group_id, "dataset", H5T_NATIVE_REAL, dataspace_id, H5P_DEFAULT, H5P_DEFAULT, H5P_DEFAULT)

  ! Write data to the dataset
  status = H5Dwrite(dataset_id, H5T_NATIVE_REAL, H5S_ALL, H5S_ALL, H5P_DEFAULT, data)

  ! Close the dataset, dataspace, group, and file
  status = H5Dclose(dataset_id)
  status = H5Sclose(dataspace_id)
  status = H5Gclose(group_id)
  status = H5Fclose(file_id)

end program create_hdf5_file

In this example, we create a new HDF5 file named "example.h5" and create a group called "group" within the file. We then create a dataset called "dataset" within the group and write some data to it.

Example 2: Reading Data from an HDF5 File

Here's an example of how to read data from an existing HDF5 file using the HDF5 Fortran API:

program read_hdf5_file
  use hdf5
  implicit none

  integer(hid_t) :: file_id, dataset_id, dataspace_id, status
  integer(hsize_t) :: dims(2)
  real :: data(10, 10)

  ! Open an existing HDF5 file
  status = H5Fopen("example.h5", H5F_ACC_RDONLY, file_id)

  ! Open the dataset within the file
  dataset_id = H5Dopen(file_id, "/group/dataset", H5P_DEFAULT)

  ! Get the dataspace of the dataset
  dataspace_id = H5Dget_space(dataset_id)

  ! Get the dimensions of the dataset
  status = H5Sget_simple_extent_dims(dataspace_id, dims, [])

  ! Read data from the dataset
  status = H5Dread(dataset_id, H5T_NATIVE_REAL, H5S_ALL, H5S_ALL, H5P_DEFAULT, data)

  ! Close the dataset, dataspace, and file
  status = H5Dclose(dataset_id)
  status = H5Sclose(dataspace_id)
  status = H5Fclose(file_id)

end program read_hdf5_file

In this example, we open an existing HDF5 file named "example.h5" and read the data from the dataset named "dataset" within the "group" group.

Conclusion

The HDF5 Fortran Framework provides a powerful and efficient way to work with HDF5 files in Fortran. Its hierarchical structure, efficient data compression, parallel I/O support, and other features make it an excellent choice for managing and storing large amounts of data. By using the HDF5 Fortran API, Fortran programmers can easily create, read, and manipulate HDF5 files in their applications.

For more information, you can refer to the official HDF5 website: https://www.hdfgroup.org/solutions/hdf5

Introduction​

History​

Features​

Examples​

Example 1: Creating an HDF5 File​

Example 2: Reading Data from an HDF5 File​

Conclusion​