A set of high level wrapper subroutines for HDF5. More...

Data Types
interface	hdf_get_dims

interface	hdf_get_mpi_int
	Get the appropriate mpi integer type. More...

interface	hdf_read_attribute
	Generic interface to read attribute. More...

interface	hdf_read_dataset
	Generic interface to read a dataset of doubles. More...

interface	hdf_read_vector_from_dataset
	Generic interface to read a vector from a dataset. More...

interface	hdf_set_even_offset

interface	hdf_write_attribute
	Generic interface to write attribute. More...

interface	hdf_write_dataset
	Generic interface to write a dataset Supported types. More...

interface	hdf_write_vector_to_dataset
	Generic interface to write a vector to dataset. More...

Functions/Subroutines
subroutine, public	hdf_set_print_messages (val_print_messages)
	Sets the value of hdf_print_messages. More...

subroutine, public	hdf_set_default_filter (filter, gzip_level, szip_pixels_per_block, szip_options)
	Sets the value of hdf_print_messages. More...

subroutine, public	hdf_exists (loc_id, obj_name, exists)
	Check if location exists. More...

subroutine, public	hdf_open_file (file_id, filename, STATUS, ACTION)
	Opens file and return identifier. More...

subroutine, public	hdf_close_file (file_id)
	Closes a hdf5 file. More...

subroutine	hdf_preset_file_attribute (file_id)
	Preset some file-level attributes. More...

subroutine, private	hdf_preset_prop ()
	Preset some properties.

subroutine, private	hdf_close_prop ()
	Preset some properties.

subroutine, public	hdf_create_group (loc_id, group_name)
	Create a new group. More...

subroutine, public	hdf_open_group (loc_id, group_name, group_id)
	Opens a group and returns the identifier. More...

subroutine, public	hdf_close_group (group_id)
	Close a group by identifier. More...

subroutine	hdf_set_even_offset_from_dataset (file_id, loc_id, dset_name, offset, new_size)
	Set even offset for read_dataset. More...

subroutine, public	hdf_get_rank (loc_id, dset_name, rank)
	Get the rank of a dataset. More...

subroutine	hdf_get_dims_4 (loc_id, dset_name, dims)
	get the dimensions of a dataset More...

subroutine	hdf_get_dims_8 (loc_id, dset_name, dims)
	get the dimensions of a dataset More...

subroutine, public	hdf_set_dims (loc_id, dset_name, dims)
	set the dimensions of a dataset

subroutine, public	hdf_create_dataset (loc_id, dset_name, dset_dims, dset_type)
	create a dataset 'dset_name' with shape 'dset_dims' of type 'dset_type' More...

Variables
integer(hid_t)	file_plist_id
	parallel file access property

integer(hid_t)	dplist_independent
	dataset access property

Detailed Description

A set of high level wrapper subroutines for HDF5.

Features:

opening and closing files
creating/opening/closing groups
get rank and dimensions of dataset
reading and writing dataset (integer, real, double)
- uses a generic interface to switch on rank and kind
writing/reading attributes (integer, real, double, string)
- uses a generic interface to switch on rank and kind

Todo:

reading and writing ( strings )
hdf_exists (h5o_exist_by_name or h5l_exists)
hdf_get_*
- hdf_get_obj_name (h5iget_name_f)
- hdf_get_obj_type (h5iget_type_f)
- hdf_get_dset_type (H5Dget_type)
- hdf_get_obj_id (not needed)
error checking,
- check dims when reading
- check dataset/attribute name when reading
- check group name when reading/writing
- stop on error vs return error flag vs global error flag

Note: I might use H5T_STD_I32LE, H5T_IEEE_F32LE, H5T_IEEE_F64LE instead of H5T_NATIVE_DOUBLE when creating a dataset. This would make the hdf5 file more portable?; should I support other integer types (ie 16 and 64)? I am not sure if hdf5 fortran supports these

Function/Subroutine Documentation

◆ hdf_set_print_messages()

subroutine, public hdf5_utils_mpi::hdf_set_print_messages ( logical, intent(in) val_print_messages )

Sets the value of hdf_print_messages.

By default, hdf_print_messages = .false. By setting it to .true., some messages are printed detailing what hdf_utils is doing.

Parameters

[in] val_print_messages new value for hdf_print_messages

◆ hdf_set_default_filter()

subroutine, public hdf5_utils_mpi::hdf_set_default_filter	(	character(len=*), intent(in)	filter,
		integer, intent(in), optional	gzip_level,
		integer, intent(in), optional	szip_pixels_per_block,
		character(len=2), intent(in), optional	szip_options
	)

Sets the value of hdf_print_messages.

Parameters

[in] filter new value for hdf_print_messages

◆ hdf_exists()

subroutine, public hdf5_utils_mpi::hdf_exists	(	integer(hid_t), intent(in)	loc_id,
		character(len=*), intent(in)	obj_name,
		logical, intent(out)	exists
	)

Check if location exists.

Also checks is intemediate paths exists in a safe way.

Parameters

[in]	loc_id	local id
[in]	obj_name	relative path to object
[out]	exists	.TRUE. if everything exists, .FALSE. otherwise

◆ hdf_open_file()

subroutine, public hdf5_utils_mpi::hdf_open_file	(	integer(hid_t), intent(out)	file_id,
		character(len=*), intent(in)	filename,
		character(len=*), intent(in), optional	STATUS,
		character(len=*), intent(in), optional	ACTION
	)

Opens file and return identifier.

Todo:

case insentive STATUS and ACTION
delete file for REPLACE case

STATUS	ACTION	Description
NEW	na	calls h5fcreate with H5F_ACC_TRUNC_F
REPLACE	na	calls h5fcreate with H5F_ACC_EXCL_F
OLD	READ	calls h5fopen with H5F_ACC_RDONLY_F
OLD	WRITE	calls h5fopen with H5F_ACC_RDWR_F
OLD	READWRITE	calls h5fopen with H5F_ACC_RDWR_F

Parameters

[out]	file_id	HDF5 id of the file
[in]	filename	the HDF5 filename
[in]	status	file status (OLD, NEW, REPLACE)
[in]	action	file action (READ, WRITE, READWRITE)

◆ hdf_close_file()

subroutine, public hdf5_utils_mpi::hdf_close_file ( integer(hid_t), intent(in) file_id )

Closes a hdf5 file.

Parameters

[in] file_id file id to be closed

◆ hdf_preset_file_attribute()

subroutine hdf5_utils_mpi::hdf_preset_file_attribute ( integer(hid_t), intent(in) file_id )

private

Preset some file-level attributes.

Parameters

[in] file_id HDF5 id of the file

◆ hdf_create_group()

subroutine, public hdf5_utils_mpi::hdf_create_group	(	integer(hid_t), intent(in)	loc_id,
		character(len=*), intent(in)	group_name
	)

Create a new group.

Parameters

[in]	loc_id	location id where to put the group
[in]	group_name	name of the group

◆ hdf_open_group()

subroutine, public hdf5_utils_mpi::hdf_open_group	(	integer(hid_t), intent(in)	loc_id,
		character(len=*), intent(in)	group_name,
		integer(hid_t), intent(out)	group_id
	)

Opens a group and returns the identifier.

Parameters

[in]	loc_id	location id where to put the group
[in]	group_name	name of the group
[out]	group_id	id for the group

◆ hdf_close_group()

subroutine, public hdf5_utils_mpi::hdf_close_group ( integer(hid_t), intent(in) group_id )

Close a group by identifier.

Parameters

[in] group_id id for the group

◆ hdf_set_even_offset_from_dataset()

subroutine hdf5_utils_mpi::hdf_set_even_offset_from_dataset	(	integer(hid_t), intent(in)	file_id,
		integer(hid_t), intent(in)	loc_id,
		character(len=*), intent(in)	dset_name,
		integer, dimension(:), intent(inout)	offset,
		integer, intent(in), optional	new_size
	)

private

Set even offset for read_dataset.

Parameters

[in]	file_id	file id created with hdf_open_file
[in]	loc_id	location id created with hdf_open_group or file_id
[in]	dset_name	dataset name
[in,out]	offset	new offset to read the dataset appropriately
[in]	new_size	(optional) new size of data on this processor, used only for consistency check

◆ hdf_get_rank()

subroutine, public hdf5_utils_mpi::hdf_get_rank	(	integer(hid_t), intent(in)	loc_id,
		character(len=*), intent(in)	dset_name,
		integer, intent(out)	rank
	)

Get the rank of a dataset.

Parameters

[in]	loc_id	location id
[in]	dset_name	dataset name
[out]	rank	rank of the dataset

◆ hdf_get_dims_4()

subroutine hdf5_utils_mpi::hdf_get_dims_4	(	integer(hid_t), intent(in)	loc_id,
		character(len=*), intent(in)	dset_name,
		integer, dimension(:), intent(out)	dims
	)

private

get the dimensions of a dataset

Parameters

[in]	loc_id	location id
[in]	dset_name	name of dataset
[out]	dims	dimensions of the dataset

◆ hdf_get_dims_8()

subroutine hdf5_utils_mpi::hdf_get_dims_8	(	integer(hid_t), intent(in)	loc_id,
		character(len=*), intent(in)	dset_name,
		integer(hsize_t), dimension(:), intent(out)	dims
	)

private

get the dimensions of a dataset

Parameters

[in]	loc_id	location id
[in]	dset_name	name of dataset
[out]	dims	dimensions of the dataset

◆ hdf_create_dataset()

subroutine, public hdf5_utils_mpi::hdf_create_dataset	(	integer(hid_t), intent(in)	loc_id,
		character(len=*), intent(in)	dset_name,
		integer, dimension(:), intent(in)	dset_dims,
		character(len=*), intent(in)	dset_type
	)

create a dataset 'dset_name' with shape 'dset_dims' of type 'dset_type'

Parameters

[in]	loc_id	local id in file
[in]	dset_name	name of dataset
[in]	dset_dims	dimensions of the dataset
[in]	dset_type	type of dataset (integer or double)

Data Types

Functions/Subroutines

Variables

Detailed Description

Function/Subroutine Documentation

◆ hdf_set_print_messages()

◆ hdf_set_default_filter()

◆ hdf_exists()

◆ hdf_open_file()

◆ hdf_close_file()

◆ hdf_preset_file_attribute()

◆ hdf_create_group()

◆ hdf_open_group()

◆ hdf_close_group()

◆ hdf_set_even_offset_from_dataset()

◆ hdf_get_rank()

◆ hdf_get_dims_4()

◆ hdf_get_dims_8()

◆ hdf_create_dataset()