legate.io.hdf5.to_file#
- legate.io.hdf5.to_file(LogicalArray array, path, str dataset_name) void#
Write a LogicalArray to disk using HDF5.
If
pathalready exists at the time of writing, the file will be overwritten.pathmay be absolute or relative. If it is relative, it will be written relative to the current working directory at the time of this function call.pathmay not fully exist at the time of this function call. Any missing directories are created (with the same permissions and properties of the current process) before tasks are launched. However, no protection is provided if those directories are later deleted before the task executes - the tasks assume these directories exist when they execute.arraymust not be unbound.- Parameters:
array (LogicalArray) – The array to serialize.
path (Pathlike) – Path to write to.
dataset_name (str) – The name of the data set to store the array under.
- Raises:
ValueError – If
pathwould not be a valid path name, for example if it is a directory name. Generally speaking, it should be in the form/path/to/file.h5.