d3396a7953
Bring object/dataset/group/named datatype features from revise_chunks
branch to trunk. Also CMake support for h5format_convert and a bunch of
misc. cleanups.
Tested on:
MacOSX/64 10.11.5 (amazon) w/serial, parallel & production
(h5committest forthcoming)
27 lines
1.1 KiB
Plaintext
27 lines
1.1 KiB
Plaintext
usage: h5format_convert [OPTIONS] file_name
|
|
OPTIONS
|
|
-h, --help Print a usage message and exit
|
|
-V, --version Print version number and exit
|
|
-v, --verbose Turn on verbose mode
|
|
-d dname, --dname=dataset_name Pathname for the dataset
|
|
-n, --noop Perform all the steps except the actual conversion
|
|
|
|
Examples of use:
|
|
|
|
h5format_convert -d /group/dataset file_name
|
|
Convert the dataset </group/dataset> in the HDF5 file <file_name>:
|
|
a. chunked dataset: convert the chunk indexing type to version 1 B-tree
|
|
b. compact/contiguous dataset: downgrade the layout version to 3
|
|
c. virtual dataset: no action
|
|
|
|
h5format_convert file_name
|
|
Convert all datasets in the HDF5 file <file_name>:
|
|
a. chunked dataset: convert the chunk indexing type to version 1 B-tree
|
|
b. compact/contiguous dataset: downgrade the layout version to 3
|
|
c. virtual dataset: no action
|
|
|
|
h5format_convert -n -d /group/dataset file_name
|
|
Go through all the steps except the actual conversion when
|
|
converting the dataset </group/dataset> in the HDF5 file <file_name>.
|
|
h5format_convert error: No dataset name
|