Groups

HDF5.open_group — Function
open_group(parent::Union{File,Group}, path::AbstractString; properties...)

Open an existing Group at path under the parent object.

Optional keyword arguments include any keywords that that belong to GroupAccessProperties.

source
HDF5.create_external — Function
create_external(source::Union{HDF5.File, HDF5.Group}, source_relpath, target_filename, target_path;
                lcpl_id=HDF5.API.H5P_DEFAULT, lapl_id=HDF5.H5P.DEFAULT)

Create an external link such that source[source_relpath] points to target_path within the file with path target_filename.

See also

API.h5l_create_external

source

AbstractDict interface

Group (like File) is a subtype of AbstractDict{String,Any}, via the shared HDF5.H5DataStore supertype. This means a Group supports the standard dict interface: keys, haskey, getindex/setindex!, get, get!, length, isempty, iteration (for (name, obj) in group, yielding name => obj pairs), pairs, values, and delete!. In particular, this enables native Julia REPL tab-completion of immediate member names, e.g. group["<tab>"]. This only completes a single level at a time (Julia's built-in dict-key completion can't decompose a /-joined path string) – for full multi-level path completion (group["some/nested/<tab>"]), install the optional companion package HDF5TreeCompletion.

A few dict operations are intentionally restricted, since their generic AbstractDict fallback would silently do something unrelated to a real HDF5 operation: copy(group) and empty(group) throw an informative error (use copy_object to copy an HDF5 object). ==/hash remain identity-based, not content-based.