Miscellaneous Example Programs

These examples resulted from working with users, and have only been tried on one or two platforms. They have NOT been fully tested.

Most of these examples are in C and were created with HDF5 1.6 or earlier.
They can be compiled with the 1.8 library* by using the option:

    -D H5_USE_16_API 

  *Unless your HDF5 1.8 library was configured with the --disable-deprecated-symbols option.

Contents:

• By API:

  Attributes	Datasets	Dataspaces	Datatypes	File
  Groups	Identifiers	Properties	References

• FORTRAN Specific
• C++ Specific - See Datatypes (Compound, String), Groups
• Parallel Specific (C)
• C#

• Datatypes - Attributes with different datatypes
• Calls H5Aiterate
• Calls H5Aiterate, where one attribute is a string array
• attrvstr.c - Shows how to read variable length string attribute (space is allocated by HDF5, but freed by user).

• Datatypes - Datasets with different datatypes

• Dynamic arrays for reading/writing data:
  • h5_readdyn.c
  • h5_writedyn.c
  • 3d.c (C only)

• Moving a Dataset

• Read in data in exponential form that is of type double, and write out to an HDF5 file:
  crtc64.c
  coordinates.txt - Text file read in

• Reading in data from text file (originally an HDF 4.x raster image, dumped with hdp), and write that out as an HDF5 file:
  h5_ri.c
  storm110.txt - Text file

• Read in floating point data from a text file and write that out as an HDF5 file:
  h5float.c
  float.txt - Text file

• Split Files: create a dataset in which the data is in a separate directory from the metadata:
  h5dirsplit.c

• Unlimited Dimension Dataset (loop by slabs):
  appslab.c
  uslab.c - Similar to appslab.c, but uses H5Pset_buffer

• Write to Finite Dimension Dataset by slabs:   wrtslab.c
  

• Writing to 4D dataset by slab and how can have performance issue:   h5pmem.c
  

• How to use selections in memory and file dataspaces (see H5Dwrite in RM):   h5_write_sel.c
• These are example programs that you can use to experiment with hyperslab selections for simple 1 and 2-Dimensional Datasets:
  • crt1d.c
  • crt2d.c

• Examples of Creating Datasets with many datatypes (atomic, enumerated, compound, string, bitfield, opaque):
  Sample Programs

• Examples of Using H5Tconvert:
  cvrt.c - one value
  cvrt2.c - two values

• Using Different Datatypes with Datasets/Attributes:
  • Array Datatypes:
    h5_array_compound.c - Create and write a dataset with a compound datatype. One of the fields has an array datatype.
    h5_array_type0.c - Create and write a dataset with an array datatype. Information about the array datatype is read back and displayed.

  • Bitfields and Opaque Datatypes

  • Boolean

  • Compound Datatypes:
    
    • Write in a loop (finite size): cmplp.c
      
    • Get class/type of each member of compound datatype: cmpclass.c
      
    • Unlimited dataset:
      h5_cmpunl.c
      cmpunlslab.c - loop through, extending dataset and writing data by slabs
      cmploop.c - compound datatype contains 2 strings
      cmp5loop.c - create 5 compound datasets (with strings) in a loop
      
    • Nested:
      • Compound Datatype (F90):   h5nestcmpd.f90
      • h5neststr.c
      • Shows how to do something like cmpd.cmpdarr[3].submember:   h5cmparr.c
      • Shows how to do something like cmpd.cmpmember.submember[5]:   h5nestc.c
    • Update array field (H5Tarray_create):  editcf.c
      
    • Update one field only in a dataset:   chg1fld.c
      
    • Update one field in every element of a dataset:   chgfield.c
      
    • With Reference:   cmpref.cpp, cmpref.h5
      
    • With String:   h5cmpstr.c
      
    • With Variable Length String:   cmpvstr.c
      Different values in string fields: cmpvlstrar.c
      Edit one string field: cmpvlstred.c     cmpvl.c
      Using H5Tvlen_create (string): cmpvlcrt.c
      
    • Using H5Tvlen_create (float): cmpvflt.c
      
    • With Enum:   h5cmpenum.c
      cenum.c - 2D array
      cenumapp.c - Append to a 2D array
      

  • Enum
    Compound Enum
    Enum with H5T_STD_U8BE datatype

  • Short Datatype (H5T_STD_I16BE) - C
    Storing 16-bit integers in F90

  • Strings:
    • Compound String
      
    • String Array - Attribute
    • String Array -Dataset
    • 1D Unlimited String Array - Dataset
    • String Attribute (C)- grpatt.c
      
    • String Attribute (C) - attrs.c
      
    • String Attribute (C++)
      String Attribute (Java) - This example requires that you first run the tutorial example which creates the dataset, dset.h5:
                http://hdf.ncsa.uiuc.edu/HDF5/doc/Tutor/crtdat.html
    • String Dataset (C)
    • String Dataset (F90)
    • String Dataset (Java)
    • Unlimited Variable Length String (uvlstr.c)
    • Variable Length String (fixed) - write by slabs (vlslab.c)
    • Variable String Attribute - Create an attribute with a variable string datatype
    • Variable Length String Array Attribute
    • Creating/Reading Variable Length String Dataset - C++ (varlen.cpp)

  • Time (H5T_TIME)

  • VL Datatypes:
    h5vlchar.c - Variable length native char example
    vlf.f90 - FORTRAN 90 example of creating dataset with variable length datatype
    h5_vltypes_ug.c - Create a dataset with a VL type (H5T_NATIVE_INT), using user defined memory management functions.
    
    [PLEASE BE AWARE that this code does not work on all platforms. It is just for showing how to use the user-defined memory management functions.]
    h5_vltypes_tut.c - Create a dataset with a VL datatype (H5T_NATIVE_INT), using the default memory management functions.

File:

• File Family
• Split File into metadata and data
• Check if objects are open

• h5group.cpp - Iterating over a group using C++
• Iterating - Call H5Giterate 100 times, returning a value of 1 in each operator call
• Scan file and print info about objects:   h5_info.c
• Moving objects
• Check if group exists
• Iterate over groups using C++ API:   h5grpnam.cpp

• Calls H5Iget_name and H5Iget_type
• Check if objects are open

• h5close.c - Using the File Close Degree Property
• h5ublock.c - Using the "User Block"

• h5ref.c - How to get information about a dereferenced object.
• h5ref2d.c - call H5Sget_select_hyper_blocklist for 2D dataset
• h5ref3d.c - call H5Sget_select_hyper_blocklist for 3D dataset

• Example code from the Belgian Institute for Space Astronomy of reading HDF5 files with Fortran.
• There are a few other example programs under Datatypes. Search for "F90".

• If you are having problems building/using Parallel HDF5, here are some handy sample programs that you can run to be sure that MPI I/O is installed properly on your machine for use with HDF5:

    Sample MPI programs
  

• The following examples show how to write data collectively and independently when one process doesn't have data or does not need to write data.
  • coll_test.c - Uses H5Sset_none to tell H5Dwrite call that there will be no data. 4-th process HAS to participate since we are in a collective mode.

  • ind_test.c - Specifies which process writes data. H5Dwrite is not called by the 4-th process at all in this case; this approach will work only when independent mode is used.

Last Modified: May 5, 2008