hdf5.group {rhdf5} | R Documentation |
In the HDF5 hierarchy there are files, groups and datasets. Files contain groups which contain datasets. The datasets are the actual data the other two levels are organizational.
hdf5.group(group, name) x[i] x[i,j, ...] x[i,j, ..., drop=TRUE] x[[i]] hdf5.group.rm(group, name)
x |
An object of class hdf5.group . |
i |
An index into the object. |
j |
An index into the object. |
... |
Optional arguments. |
drop |
A logical value, if TRUE the result may drop
dimensions. |
group |
An instance of hdf5.file or
hdf5.group . |
name |
A character vector naming the new group. |
A em{group} in the HDF5 since is merely an organizational structure. The basic unit is an HDF5 file, which contains, groups and groups contain datasets. It is possible for a dataset to belong to more than one group.
The single-bracket subset operator,
[
, is not implemented. It should return a new group, but since
there is no way to name that group, it is not clear what to do.
hdf.group.rm
removes (or unlinks) the HDF5 object named
name
from the HDF5 object group
. Note that caution
should be used since it is possible to make data inaccessible by
removing all links to it.
The [[
and $
operators return the named element of a
group; which can be either an hdf5.group
or an
hdf5.dataset
.
hdf.group.rm
returns its first argument (group
),
invisibly.
R. Gentleman and B. Ellis
## Not run: mad <- hdf5.open("microarray.h5") hdf5.group(mad,"chip1") hdf5.group(mad$chip1,"group") hdf5.group.rm(mad$chip1,"group") ## End(Not run)