Struct typed_store::rocks::DBBatch

pub struct DBBatch { /* private fields */ }

Provides a mutable struct to form a collection of database write operations, and execute them.

Batching write and delete operations is faster than performing them one by one and ensures their atomicity, ie. they are all written or none is. This is also true of operations across column families in the same database.

Serializations / Deserialization, and naming of column families is performed by passing a DBMap<K,V> with each operation.

use typed_store::rocks::*;
use tempfile::tempdir;
use typed_store::Map;
use typed_store::metrics::DBMetrics;
use prometheus::Registry;
use core::fmt::Error;
use std::sync::Arc;

#[tokio::main]
async fn main() -> Result<(), Error> {
let rocks = open_cf(tempfile::tempdir().unwrap(), None, MetricConf::default(), &["First_CF", "Second_CF"]).unwrap();

let db_cf_1 = DBMap::reopen(&rocks, Some("First_CF"), &ReadWriteOptions::default())
    .expect("Failed to open storage");
let keys_vals_1 = (1..100).map(|i| (i, i.to_string()));

let db_cf_2 = DBMap::reopen(&rocks, Some("Second_CF"), &ReadWriteOptions::default())
    .expect("Failed to open storage");
let keys_vals_2 = (1000..1100).map(|i| (i, i.to_string()));

let mut batch = db_cf_1.batch();
batch
    .insert_batch(&db_cf_1, keys_vals_1.clone())
    .expect("Failed to batch insert")
    .insert_batch(&db_cf_2, keys_vals_2.clone())
    .expect("Failed to batch insert");

let _ = batch.write().expect("Failed to execute batch");
for (k, v) in keys_vals_1 {
    let val = db_cf_1.get(&k).expect("Failed to get inserted key");
    assert_eq!(Some(v), val);
}

for (k, v) in keys_vals_2 {
    let val = db_cf_2.get(&k).expect("Failed to get inserted key");
    assert_eq!(Some(v), val);
}
Ok(())
}

Implementations

impl DBBatch

pub fn new( dbref: &Arc<RocksDB>, batch: RocksDBBatch, opts: WriteOptions, db_metrics: &Arc<DBMetrics>, write_sample_interval: &SamplingInterval ) -> Self

Create a new batch associated with a DB reference.

Use open_cf to get the DB reference or an existing open database.

pub fn write(self) -> Result<(), TypedStoreError>

Consume the batch and write its operations to the database

impl DBBatch

pub fn delete_batch<J: Borrow<K>, K: Serialize, V>( &mut self, db: &DBMap<K, V>, purged_vals: impl IntoIterator<Item = J> ) -> Result<(), TypedStoreError>

pub fn schedule_delete_range<K: Serialize, V>( &mut self, db: &DBMap<K, V>, from: &K, to: &K ) -> Result<(), TypedStoreError>

Deletes a range of keys between from (inclusive) and to (non-inclusive) by writing a range delete tombstone in the db map If the DBMap is configured with ignore_range_deletions set to false, the effect of this write will be visible immediately i.e. you won’t see old values when you do a lookup or scan. But if it is configured with ignore_range_deletions set to true, the old value are visible until compaction actually deletes them which will happen sometime after. By default ignore_range_deletions is set to true on a DBMap (unless it is overriden in the config), so please use this function with caution

pub fn insert_batch<J: Borrow<K>, K: Serialize, U: Borrow<V>, V: Serialize>( &mut self, db: &DBMap<K, V>, new_vals: impl IntoIterator<Item = (J, U)> ) -> Result<&mut Self, TypedStoreError>

inserts a range of (key, value) pairs given as an iterator

pub fn merge_batch<J: Borrow<K>, K: Serialize, U: Borrow<V>, V: Serialize>( &mut self, db: &DBMap<K, V>, new_vals: impl IntoIterator<Item = (J, U)> ) -> Result<&mut Self, TypedStoreError>

merges a range of (key, value) pairs given as an iterator

pub fn partial_merge_batch<J: Borrow<K>, K: Serialize, V: Serialize, B: AsRef<[u8]>>( &mut self, db: &DBMap<K, V>, new_vals: impl IntoIterator<Item = (J, B)> ) -> Result<&mut Self, TypedStoreError>

similar to merge_batch but allows merge with partial values