Question

I am trying to load an hdf5 into R and running into some problems. Here are the steps I took to configure my environment:

• R 2.10.0 (x64) on Mac OS X 10.6
• hdf5 1.8.3 installed via macports
• hdf5_1.6.9.tar.gz from CRAN

I suspect the problem I am having relates to incompatibilities in my version of HDF5 and the one the R module expects. For completeness here is how I installed the R module:

R CMD INSTALL –configure-vars=’CPPFLAGS=-I/opt/local/include’ –configure-args=’–with-hdf5=/opt/local’ hdf5_1.6.9.tar.gz

This builds fine. The library seems to load without issue, but no data is returned when I try to load a file:

library(hdf5)

hdf5load(“test.h5”)

NULL

Yet,

osx:data scott$ h5dump test.h5
HDF5 “test.h5” {
GROUP “/” {
DATASET “dset” {
DATATYPE H5T_STD_I32LE
DATASPACE SIMPLE { ( 31 ) / ( 31 ) }
DATA {
(0): 1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8192,
(14): 16384, 32768, 65536, 131072, 262144, 524288, 1048576, 2097152,
(22): 4194304, 8388608, 16777216, 33554432, 67108864, 134217728,
(28): 268435456, 536870912, 1073741824
}
}
}
}

Any thoughts?

Thanks in advance.

Answer 1

Use the verbosity argument in hdfload() and check your environment using ls(). hdfload() has a side effect, it loads the tables into the current environment, NULL is the default return value. If you use the argument load=TRUE the objects are returned as the components of a named list:

• load: A logical value. If ‘FALSE’,
  the objects are returned as the
  components of a named list. If ‘TRUE’ (the default), the
  objects are loaded as individual variables with their own
  names – the function returns nothing in this case.

On my Mac everything works well with the same setup.

R> rm(list=ls())
R> ls()
character(0)
R> hdf5load("TestHDF5.h5", verbosity=3)
hdf5_global_verbosity=3 load=1
Processing object: Dataset0 ...... its a dataset...Dataset has ID335544326
Dataset has tid 201326902
Dataset has space id 268435467
Dataset has rank 1
Dataset has dims/maxdims: 1 / 1 
Allocating vector with rank=1 dim=1
calling vector_io. Hangs here with big datsets
Setting buffer size in plist
About to read with bufsize = 50
in string_ref: count=1, size=25 srcbf=25
leaving string_ref
 Done read
in vector_io: permuting
in vector_io: tidying
Phew. Done it. calling iinfo->add
Rank > 1 or not VECSXP
Calling  hdf5_load_attributes 
back from  hdf5_load_attributes 
...Finished dataset 
Processing object: Table0 ...... its a dataset...Dataset has ID335544327
Dataset has tid 201326906
Dataset has space id 268435468
Dataset has rank 1
Dataset has dims/maxdims: 1 / 1 
Dataset has type = VECSXP and rank 1
Reading...
....done
in string_ref: count=1, size=25 srcbf=25
leaving string_ref
...Finished dataset 
NULL
R> ls()
[1] "Dataset0" "Table0"