#database #orm #mysql #postgres #sqlite


The Rust SQL Toolkit and ORM Library. An async, pure Rust SQL crate featuring compile-time Dynamic SQL

160 stable releases

4.0.49 Jan 3, 2023
4.0.47 Dec 30, 2022
4.0.43 Nov 17, 2022
3.1.16 Jul 28, 2022
1.0.7 Feb 29, 2020

#14 in Database interfaces

Download history 371/week @ 2022-10-15 589/week @ 2022-10-22 628/week @ 2022-10-29 656/week @ 2022-11-05 516/week @ 2022-11-12 455/week @ 2022-11-19 296/week @ 2022-11-26 349/week @ 2022-12-03 380/week @ 2022-12-10 365/week @ 2022-12-17 525/week @ 2022-12-24 313/week @ 2022-12-31 278/week @ 2023-01-07 187/week @ 2023-01-14 410/week @ 2023-01-21 411/week @ 2023-01-28

1,344 downloads per month
Used in 15 crates (10 directly)


10K SLoC

WebSite | Showcase

Build Status doc.rs unsafe forbidden GitHub release discussions

A async, pure Rust SQL Toolkit,compile-time Dynamic SQL,Compile Time ORM.

It is an ORM, a small compiler, a dynamic SQL languages

  • Non-invasive API design.You don't need to change the current code
  • Compatible with most mybatis3 syntax
  • No Runtimes,No Garbage Collection,High performance, Based on Future/Tokio
  • Zero cost Dynamic SQL, implemented using (proc-macro,compile-time,Cow(Reduce unnecessary cloning)) techniques。 don't need ONGL engine(mybatis)
  • JDBC-like driver design, driver use cargo.toml dependency and Box<dyn Driver> separation
  • All database drivers supported #{arg}, ${arg},? placeholder(pg/mssql auto processing '?' to '$1' and '@P1')
  • Dynamic SQL(Write code freely in SQL),pagination, py_sql query lang and html_sql(Inspired Mybatis).
  • Dynamic configuration connection pool(Based on the deadpool)
  • Supports logging, customizable logging based on log crate
  • 100% Safe Rust with #![forbid(unsafe_code)] enabled
  • Support use Trait System Add py_sql/ html_sql functions.see
  • abs_admin project an complete background user management system( Vue.js+rbatis+actix-web)

Thanks to SQLX, deadpool, Tiberius, MyBatis,xorm and so on reference design or code implementation. release of V4.0 is Inspired and supported by these frameworks


  • this bench test is MockTable,MockDriver,MockConnection to Assume that the network I/O time is 0
  • run code rbatis.fetch_decode::<Vec<i32>>("", vec![]).await; on benches bench_raw()
  • run code MockTable::insert(&mut rbatis.clone(),&t).await; on benches bench_insert()
  • run code MockTable::select_all(&mut rbatis.clone()).await.unwrap(); on benches bench_select()
---- bench_raw stdout ----
Time: 84.6463ms ,each:846 ns/op
QPS: 1181286 QPS/s

---- bench_insert stdout ----(macos,cpu-M1Max)
Time: 378.186333ms ,each:3781 ns/op
QPS: 264418 QPS/s

---- bench_select stdout ----(macos,cpu-M1Max)
Time: 112.927916ms ,each:1129 ns/op
QPS: 885486 QPS/s

Supported data structures

data structure is supported
i32,i64,f32,f64,bool,String...more rust type
rbatis::plugin::page::{Page, PageRequest}
any serde type
driver type on package (rdbc-mysql/types,rbdc-pg/types,rbdc-sqlite/types)

Supported database driver

database(crates.io) github_link
Mysql rbatis/rbdc-mysql
Postgres rbatis/rbdc-pg
Sqlite rbatis/rbdc-sqlite
Mssql rbatis/rbdc-mssql
MariaDB rbatis/rbdc-mysql
TiDB rbatis/rbdc-mysql
CockroachDB rbatis/rbdc-pg
Oracle chenpengfan/rbdc-oracle

Supported OS/Platforms by Workflows CI

platform is supported
Linux(unbutu laster***)

Supported Web Frameworks

  • any web Frameworks just like actix-web,axum,hyper*,rocket,tide,warp,salvo...... and more
Quick example: QueryWrapper and common usages (see example/crud_test.rs for details)
  • Rust v1.63+ later

  • Cargo.toml

tokio = { version = "1", features = ["full"] }
log = "0.4"
fast_log = "1.5"
serde = { version = "1", features = ["derive"] }
rbs = { version = "0.1"}
rbatis = { version = "4.0"}
rbdc-sqlite = { version = "0.1" }
#...and more driver
//#[macro_use] define in 'root crate' or 'mod.rs' or 'main.rs'
extern crate rbatis;
extern crate rbdc;
use rbatis::rbdc::datetime::FastDateTime;

#[derive(Clone, Debug, Serialize, Deserialize)]
pub struct BizActivity {
    pub id: Option<String>,
    pub name: Option<String>,
    pub pc_link: Option<String>,
    pub h5_link: Option<String>,
    pub pc_banner_img: Option<String>,
    pub h5_banner_img: Option<String>,
    pub sort: Option<String>,
    pub status: Option<i32>,
    pub remark: Option<String>,
    pub create_time: Option<FastDateTime>,
    pub version: Option<i64>,
    pub delete_flag: Option<i32>,
crud!(BizActivity{});//crud = insert+select_by_column+update_by_column+delete_by_column

impl_select!(BizActivity{select_all_by_id(id:&str,name:&str) => "`where id = #{id} and name = #{name}`"});
impl_select!(BizActivity{select_by_id(id:String) -> Option => "`where id = #{id} limit 1`"});
impl_update!(BizActivity{update_by_name(name:&str) => "`where id = 1`"});
impl_delete!(BizActivity {delete_by_name(name:&str) => "`where name= '2'`"});
impl_select_page!(BizActivity{select_page(name:&str) => "`where name != #{name}`"});

async fn main() {
    /// enable log crate to show sql logs
    fast_log::init(fast_log::Config::new().console()).expect("rbatis init fail");
    /// initialize rbatis. also you can call rb.clone(). this is  an Arc point
    let rb = Rbatis::new();
    /// connect to database  
    // sqlite 
    rb.init(SqliteDriver {}, "sqlite://target/sqlite.db").unwrap();
    // mysql 
    // rb.init(MysqlDriver{},"mysql://root:123456@localhost:3306/test").unwrap();
    // postgresql 
    // rb.init(PgDriver{},"postgres://postgres:123456@localhost:5432/postgres").unwrap();
    // mssql/sqlserver
    // rb.init(MssqlDriver{},"jdbc:sqlserver://localhost:1433;User=SA;Password={TestPass!123456};Database=test").unwrap();

    let activity = BizActivity {
        id: Some("2".into()),
        name: Some("2".into()),
        pc_link: Some("2".into()),
        h5_link: Some("2".into()),
        pc_banner_img: None,
        h5_banner_img: None,
        sort: None,
        status: Some(2),
        remark: Some("2".into()),
        create_time: Some(FastDateTime::now()),
        version: Some(1),
        delete_flag: Some(1),
    let data = BizActivity::insert(&mut rb, &activity).await;
    println!("insert = {:?}", data);

    let data = BizActivity::select_all_by_id(&mut rb, "1", "1").await;
    println!("select_all_by_id = {:?}", data);

    let data = BizActivity::select_by_id(&mut rb, "1".to_string()).await;
    println!("select_by_id = {:?}", data);

    let data = BizActivity::update_by_column(&mut rb, &activity, "id").await;
    println!("update_by_column = {:?}", data);

    let data = BizActivity::update_by_name(&mut rb, &activity, "test").await;
    println!("update_by_name = {:?}", data);

    let data = BizActivity::delete_by_column(&mut rb, "id", &"2".into()).await;
    println!("delete_by_column = {:?}", data);

    let data = BizActivity::delete_by_name(&mut rb, "2").await;
    println!("delete_by_column = {:?}", data);

    let data = BizActivity::select_page(&mut rb, &PageRequest::new(1, 10), "2").await;
    println!("select_page = {:?}", data);
///...more usage,see crud.rs
  • raw-sql
pub async fn main() {
    use rbatis::Rbatis;
    use rbdc_sqlite::driver::SqliteDriver;
    #[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
    pub struct BizActivity {
        pub id: Option<String>,
        pub name: Option<String>,
    fast_log::init(fast_log::Config::new().console()).expect("rbatis init fail");
    let rb = Rbatis::new();
    rb.init(SqliteDriver {}, "sqlite://target/sqlite.db").unwrap();
    let table: Option<BizActivity> = rb
        .fetch_decode("select * from biz_activity limit ?", vec![rbs::to_value!(1)])
    let count: u64 = rb
        .fetch_decode("select count(1) as count from biz_activity", vec![])
    println!(">>>>> table={:?}", table);
    println!(">>>>> count={}", count);


  • Important update (pysql removes runtime, directly compiles to static rust code) This means that the performance of SQL generated using py_sql,html_sql is roughly similar to that of handwritten code.

Because of the compile time, the annotations need to declare the database type to be used.

    #[py_sql("select * from biz_activity where delete_flag = 0
                  if name != '':
                    `and name=#{name}`")]
async fn py_sql_tx(rb: &Rbatis, tx_id: &String, name: &str) -> Vec<BizActivity> { impled!() }
  • Added html_sql support, a form of organization similar to MyBatis, to facilitate migration of Java systems to Rust( Note that it is also compiled as Rust code at build time and performs close to handwritten code) this is very faster

Because of the compile time, the annotations need to declare the database type to be used

    <select id="select_by_condition">
        `select * from biz_activity where `
        <if test="name != ''">
            name like #{name}
    ///select page must have  '?:&PageRequest' arg and return 'Page<?>'
async fn select_by_condition(rb: &mut dyn Executor, page_req: &PageRequest, name: &str) -> Page<BizActivity> { impled!() }
use once_cell::sync::Lazy;

pub static RB: Lazy<Rbatis> = Lazy::new(|| Rbatis::new());

/// Macro generates execution logic based on method definition, similar to @select dynamic SQL of Java/Mybatis
/// RB is the name referenced locally by Rbatis, for example DAO ::RB, com:: XXX ::RB... Can be
/// The second parameter is the standard driver SQL. Note that the corresponding database parameter mysql is? , pg is $1...
/// macro auto edit method to  'pub async fn select(name: &str) -> rbatis::core::Result<BizActivity> {}'
#[sql("select * from biz_activity where id = ?")]
pub async fn select(rb: &Rbatis, name: &str) -> BizActivity {}
//or: pub async fn select(name: &str) -> rbatis::core::Result<BizActivity> {}

pub async fn test_macro() {
    fast_log::init(fast_log::Config::new().console()).expect("rbatis init fail");
    let a = select(&RB, "1").await.unwrap();
    println!("{:?}", a);

How it works

Rely on rbatis-codegen to create the source code of the corresponding structure from the html file at compile time (with debug_mode(Cargo.toml- rbatis = { features = ["debug_mode"]}) enabled, you can observe the code-generated function), and call the generated method directly at run time. We know that compilation is generally divided into three steps, lexes, syntactic analysis, semantic analysis, and intermediate code generation. In rbatis, Lexical analysis is handled by the dependent func.rs in rbatis-codegen, which relies on syn and quote. Parsing is done by parser_html and parser_pysql in rbatis-codegen The generated syntax tree is a structure defined in the syntax_tree package in rbatis-codegen Intermediate code generation has func.rs generation function, all supported functions are defined in rbatis-codegen

What is described above occurs during the cargo build phase, which is the compilation phase of the rust procedural macro, where the code generated by rbatis-codegen is handed back to the rust compiler for LLVM compilation to produce pure machine code

So I think rbatis is Truly zero overhead dynamic SQL compile-time ORM.

Submit PR(Pull Requests)

You are welcome to submit the merge, and make sure that any functionality you add has the appropriate mock unit test function added under the test package.



  • sqlite table sync plugin(auto create table/column)
  • connection pool add more dynamically configured parameters
  • More simplified syntax

Contact/donation, or click on star rbatis

  • discussions

联系方式/捐赠,或 rbatis 点star



联系方式(添加好友请备注'rbatis') 微信群:先加微信,然后拉进群



~511K SLoC