Crates.io | qldb |
lib.rs | qldb |
version | 3.2.9 |
source | src |
created_at | 2020-11-11 16:29:54.624805 |
updated_at | 2024-05-30 12:08:22.530831 |
description | Driver for Amazon's QLDB Database implemented in pure rust. |
homepage | https://crates.io/crates/qldb |
repository | https://github.com/Couragium/qldb-rs |
max_upload_size | |
id | 311296 |
size | 118,295 |
Driver for Amazon's QLDB Database implemented in pure rust.
The driver is fairly tested and should be ready to test in real projects.
use qldb::QldbClient;
use std::collections::HashMap;
let client = QldbClient::default("rust-crate-test", 200).await?;
let mut value_to_insert = HashMap::new();
// This will insert a documents with a key "test_column"
// with the value "IonValue::String(test_value)"
value_to_insert.insert("test_column", "test_value");
client
.transaction_within(|client| async move {
client
.query("INSERT INTO TestTable VALUE ?")
.param(value_to_insert)
.execute()
.await?;
Ok(())
})
.await?;
The driver has a session pool. The second parameter in the QldbClient::default is the maximum size of the connection pool.
The pool will be auto-populated as parallel transaction are being requested until it reaches the provided maximum.
The pool uses one independent thread with a single-threaded executor (async-executor) in order to be able to spawn tasks after the session has been returned.
There is an alternative session pool that will require an spawner function to be provided. It allows to have the pool running by using the spawn function of the executor you use. We tested async-std and tokio, but others should work as well.
This pool will spawn two internal tasks handling the pool.
Use this if you want for this driver to not create a new thread.
Example with async-std:
let client = QldbClient::default_with_spawner(
"rust-crate-test",
200,
Arc::new(move |fut| {async_std::task::spawn(Box::pin(fut));})
)
.await?
Or, with tokio:
let client = QldbClient::default_with_spawner(
"rust-crate-test",
200,
Arc::new(move |fut| {tokio::spawn(Box::pin(fut));})
)
.await?
By default, both pools are available by using the methods QldbClient::default
and QldbClient::default_with_spawner
. If you don't want the pool to be available
in runtime, you can disable by removing the default features. Still, you will
need to add at least one feature to enable one pool.
This will only enable the default pool, the one that uses one thread.
qldb = { version = "3", default_features = false, features = ["internal_pool_with_thread"]}
This will only enable the alternative pool, the one that requires an spawner
qldb = { version = "3", default_features = false, features = ["internal_pool_with_spawner"]}
The library uses ion-binary-rs, which is our own, pure rust, implementation of the format. It is very well tested and ready to use in production too.
For tests you will need to have some AWS credentials in your PC (as env variables or in ~/.aws/credentials). There needs to be a QLDB database with the name "rust-crate-test" in the aws account. The tests need to be run sequentially, so in order to run the tests please run the following command:
RUST_TEST_THREADS=1 cargo test