Struct Dataset 

pub struct Dataset<'f> { /* private fields */ }

A dataset within an HDF5 file.

Implementations

impl<'f> Dataset<'f>

pub fn name(&self) -> &str

Dataset name.

pub fn address(&self) -> u64

The object header address used to parse this dataset. Useful as an opaque identifier or for NC4 data_offset.

pub fn shape(&self) -> &[u64]

Shape of the dataset (dimensions).

pub fn dtype(&self) -> &Datatype

Datatype of the dataset.

pub fn ndim(&self) -> usize

Number of dimensions.

pub fn max_dims(&self) -> Option<&[u64]>

Maximum dimension sizes, if defined. u64::MAX indicates unlimited.

pub fn chunks(&self) -> Option<Vec<u32>>

Chunk dimensions, if the dataset is chunked.

pub fn fill_value(&self) -> Option<&FillValueMessage>

Fill value, if defined.

pub fn attributes(&self) -> Vec<Attribute>

Dataset attributes.

pub fn attribute(&self, name: &str) -> Result<Attribute>

Find an attribute by name.

pub fn read_string(&self) -> Result<String>

Read a scalar string dataset or single string element.

Use Dataset::read_strings when the dataset contains multiple strings.

pub fn read_strings(&self) -> Result<Vec<String>>

Read all string elements from a string-typed dataset.

pub fn num_elements(&self) -> u64

Total number of elements in the dataset.

pub fn read_array<T: H5Type>(&self) -> Result<ArrayD<T>>

Read the entire dataset into an n-dimensional array.

pub fn read_array_parallel<T: H5Type>(&self) -> Result<ArrayD<T>>

Read the entire dataset using internal chunk-level parallelism when possible.

Non-chunked datasets fall back to read_array.

pub fn read_array_in_pool<T: H5Type>( &self, pool: &ThreadPool, ) -> Result<ArrayD<T>>

Read the entire dataset using the provided Rayon thread pool.

Non-chunked datasets fall back to read_array.

pub fn read_slice_parallel<T: H5Type>( &self, selection: &SliceInfo, ) -> Result<ArrayD<T>>

Read a hyperslab of the dataset using chunk-level parallelism when possible.

Chunked datasets decompress overlapping chunks in parallel via Rayon. Non-chunked layouts fall back to read_slice.

pub fn read_slice<T: H5Type>(&self, selection: &SliceInfo) -> Result<ArrayD<T>>

Read a hyperslab of the dataset.