Struct below_store::StoreWriter

pub struct StoreWriter { /* private fields */ }

The StoreWriter struct maintains state to put more data in the store. It keeps track of the index and data file it’s currently working on so in the common case it can just append data. When it rolls over to a new shard, it will recreate itself.

Implementations

impl StoreWriter

pub fn new<P: AsRef<Path>>( logger: Logger, path: P, compression_mode: CompressionMode, format: Format ) -> Result<Self>

Create a new StoreWriter that writes data to path directory. Data serialized with format.

If compression_mode is set, dataframes are zstd compressed, as defined by compression_mode.

pub fn new_with_timestamp<P: AsRef<Path>>( logger: Logger, path: P, timestamp: SystemTime, compression_mode: CompressionMode, format: Format ) -> Result<Self>

pub fn put(&mut self, timestamp: SystemTime, data: &DataFrame) -> Result<bool>

Store data with corresponding timestamp. Returns true if a new shard is created and data is written successfully. Errors may be returned if file operations fail.

pub fn discard_earlier(&self, timestamp: SystemTime) -> Result<()>

Discard all data earlier than timestamp

We do not modify index and data files. We just look for files which can only contain earlier data and remove them.

pub fn try_discard_until_size(&self, store_size_limit: u64) -> Result<bool>

Discard data until store size is less than limit, or there is only one shard left. Oldest shards are discarded first. Returns true on success or false if the current shard size is greater than the limit.