Struct rkv::Rkv

source ·

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

Expand description

Wrapper around an Environment (e.g. such as an LMDB or SafeMode environment).

Implementations§

source §

impl<'e, E> Rkv<E>where E: BackendEnvironment<'e>,

Static methods.

source

pub fn environment_builder() -> Bwhere B: BackendEnvironmentBuilder<'e, Environment = E>,

source

pub fn new(path: &Path) -> Result<Rkv<E>, StoreError>where B: BackendEnvironmentBuilder<'e, Environment = E>,

Return a new Rkv environment that supports up to DEFAULT_MAX_DBS open databases.

source

pub fn with_capacity( path: &Path, max_dbs: c_uint ) -> Result<Rkv<E>, StoreError>where B: BackendEnvironmentBuilder<'e, Environment = E>,

Return a new Rkv environment that supports the specified number of open databases.

source

pub fn from_builder(path: &Path, builder: B) -> Result<Rkv<E>, StoreError>where B: BackendEnvironmentBuilder<'e, Environment = E>,

Return a new Rkv environment from the provided builder.

source §

impl<'e, E> Rkv<E>where E: BackendEnvironment<'e>,

Store creation methods.

source

pub fn get_dbs(&self) -> Result<Vec<Option<String>>, StoreError>

Return all created databases.

source

pub fn open_single<'s, T>( &self, name: T, opts: StoreOptions<E::Flags> ) -> Result<SingleStore<E::Database>, StoreError>where T: Into<Option<&'s str>>,

Create or Open an existing database in (&u8 -> Single Value) mode. Note: that create=true cannot be called concurrently with other operations so if you are sure that the database exists, call this with create=false.

source

pub fn open_integer<'s, T, K>( &self, name: T, opts: StoreOptions<E::Flags> ) -> Result<IntegerStore<E::Database, K>, StoreError>where K: PrimitiveInt, T: Into<Option<&'s str>>,

Create or Open an existing database in (Integer -> Single Value) mode. Note: that create=true cannot be called concurrently with other operations so if you are sure that the database exists, call this with create=false.

source

pub fn open_multi<'s, T>( &self, name: T, opts: StoreOptions<E::Flags> ) -> Result<MultiStore<E::Database>, StoreError>where T: Into<Option<&'s str>>,

Create or Open an existing database in (&u8 -> Multiple Values) mode. Note: that create=true cannot be called concurrently with other operations so if you are sure that the database exists, call this with create=false.

source

pub fn open_multi_integer<'s, T, K>( &self, name: T, opts: StoreOptions<E::Flags> ) -> Result<MultiIntegerStore<E::Database, K>, StoreError>where K: PrimitiveInt, T: Into<Option<&'s str>>,

Create or Open an existing database in (Integer -> Multiple Values) mode. Note: that create=true cannot be called concurrently with other operations so if you are sure that the database exists, call this with create=false.

source §

impl<'e, E> Rkv<E>where E: BackendEnvironment<'e>,

Read and write accessors.

source

pub fn read<T>(&'e self) -> Result<Reader<T>, StoreError>where E: BackendEnvironment<'e, RoTransaction = T>, T: BackendRoCursorTransaction<'e, Database = E::Database>,

Create a read transaction. There can be multiple concurrent readers for an environment, up to the maximum specified by LMDB (default 126), and you can open readers while a write transaction is active.

source

pub fn write<T>(&'e self) -> Result<Writer<T>, StoreError>where E: BackendEnvironment<'e, RwTransaction = T>, T: BackendRwCursorTransaction<'e, Database = E::Database>,

Create a write transaction. There can be only one write transaction active at any given time, so trying to create a second one will block until the first is committed or aborted.

source §

impl<'e, E> Rkv<E>where E: BackendEnvironment<'e>,

Other environment methods.

source

pub fn sync(&self, force: bool) -> Result<(), StoreError>

Flush the data buffers to disk. This call is only useful, when the environment was open with either NO_SYNC, NO_META_SYNC or MAP_ASYNC (see below). The call is not valid if the environment was opened with READ_ONLY.

Data is always written to disk when transaction.commit() is called, but the operating system may keep it buffered. LMDB always flushes the OS buffers upon commit as well, unless the environment was opened with NO_SYNC or in part NO_META_SYNC.

force: if true, force a synchronous flush. Otherwise if the environment has the NO_SYNC flag set the flushes will be omitted, and with MAP_ASYNC they will be asynchronous.

source

pub fn stat(&self) -> Result<E::Stat, StoreError>

Retrieve statistics about this environment.

It includes:

Page size in bytes
B-tree depth
Number of internal (non-leaf) pages
Number of leaf pages
Number of overflow pages
Number of data entries

source

pub fn info(&self) -> Result<E::Info, StoreError>

Retrieve information about this environment.

It includes:

Map size in bytes
The last used page number
The last transaction ID
Max number of readers allowed
Number of readers in use

source

pub fn load_ratio(&self) -> Result<Option<f32>, StoreError>

Retrieve the load ratio (# of used pages / total pages) about this environment.

With the formular: (last_page_no - freelist_pages) / total_pages. A value of None means that the backend doesn’t ever need to be resized.

source

pub fn set_map_size(&self, size: usize) -> Result<(), StoreError>

Sets the size of the memory map to use for the environment.

This can be used to resize the map when the environment is already open. You can also use Rkv::environment_builder() to set the map size during the Rkv initialization.

Note:

No active transactions allowed when performing resizing in this process. It’s up to the consumer to enforce that.
The size should be a multiple of the OS page size. Any attempt to set a size smaller than the space already consumed by the environment will be silently changed to the current size of the used space.
In the multi-process case, once a process resizes the map, other processes need to either re-open the environment, or call set_map_size with size 0 to update the environment. Otherwise, new transaction creation will fail with LmdbError::MapResized.

source