Struct TransactionalTree 

pub struct TransactionalTree<E> { /* private fields */ }

A transaction that will be applied atomically to the Tree.

Implementations

impl<E> TransactionalTree<E>

where E: Encryption,

pub fn get<K: AsRef<[u8]>>( &self, key: K, ) -> Result<Option<IVec>, UnabortableTransactionError>

Get the value associated with a key

pub fn insert<K, V>( &self, key: K, value: V, ) -> Result<Option<IVec>, UnabortableTransactionError>

where K: AsRef<[u8]>, V: Into<IVec>,

Set a key to a new value

pub fn remove<K: AsRef<[u8]>>( &self, key: K, ) -> Result<Option<IVec>, UnabortableTransactionError>

Remove a key

pub fn apply_batch( &self, batch: &Batch, ) -> Result<(), UnabortableTransactionError>

Atomically apply multiple inserts and removals.

pub fn flush(&self)

Flush the database before returning from the transaction.

pub fn generate_id(&self) -> Result<u64>

Generate a monotonic ID. Not guaranteed to be contiguous or idempotent, can produce different values in the same transaction in case of conflicts. Written to disk every idgen_persist_interval operations, followed by a blocking flush. During recovery, we take the last recovered generated ID and add 2x the idgen_persist_interval to it. While persisting, if the previous persisted counter wasn’t synced to disk yet, we will do a blocking flush to fsync the latest counter, ensuring that we will never give out the same counter twice.