Struct NodeTable 

pub struct NodeTable { /* private fields */ }

NodeTable provides bidirectional mapping between Terms and NodeIds

Uses two B+Trees:

• term_to_id: Term → NodeId (for encoding)
• id_to_term: NodeId → Term (for decoding)

Prefix Compression

For IRI terms, the NodeTable automatically applies prefix compression to reduce storage overhead. Common IRI prefixes (namespaces) are stored once and referenced by ID, significantly reducing memory usage for datasets with many IRIs sharing the same namespace.

Implementations

impl NodeTable

pub fn new(buffer_pool: Arc<BufferPool>) -> Self

Create a new NodeTable with prefix compression enabled

pub fn with_compression( buffer_pool: Arc<BufferPool>, compression_enabled: bool, ) -> Self

Create a new NodeTable with configurable compression

pub fn from_roots( buffer_pool: Arc<BufferPool>, term_to_id_root: Option<PageId>, id_to_term_root: Option<PageId>, next_id: NodeId, ) -> Self

Reconstruct a NodeTable from persisted B+Tree roots and the next id to hand out (used when reopening a store from its superblock).

pub fn term_to_id_root(&self) -> Option<PageId>

Current root page of the term_to_id B+Tree (for superblock persistence).

pub fn id_to_term_root(&self) -> Option<PageId>

Current root page of the id_to_term B+Tree (for superblock persistence).

pub fn next_id_value(&self) -> NodeId

Next dictionary NodeId to be allocated (for superblock persistence).

pub fn compression_stats(&self) -> CompressionStats

Get prefix compression statistics

pub fn get_or_create(&self, term: &Term) -> Result<NodeId>

Get or create a NodeId for a term

If the term already exists, returns its NodeId. Otherwise, assigns a new NodeId and stores the mapping.

Prefix Compression

When prefix compression is enabled and the term is an IRI, the compressor tracks the IRI prefix for statistics. Future versions will use this for actual storage optimization.

pub fn get_id(&self, term: &Term) -> Result<Option<NodeId>>

Get NodeId for a term (returns None if not found)

pub fn get_term(&self, id: NodeId) -> Result<Option<Term>>

Get Term for a NodeId (returns None if not found)

pub fn size(&self) -> u64

Get the total number of terms stored

pub fn contains(&self, term: &Term) -> Result<bool>

Check if a term exists in the dictionary

pub fn contains_id(&self, id: NodeId) -> Result<bool>

Check if a NodeId exists in the dictionary