Module sqlx_migrator::migrator

Migrator module

It contains common enum and trait for implementing migrator for sqlx supported database

It also provides its own struct Migrator which supports Any, Postgres, Sqlite and Mysql database

Example

Create own custom Migrator which only supports postgres and uses own unique table name instead of default table name

use sqlx::{Pool, Postgres};
use sqlx_migrator::error::Error;
use sqlx_migrator::migration::{AppliedMigrationSqlRow, Migration};
use sqlx_migrator::migrator::{DatabaseOperation, Info, Migrate};

#[derive(Default)]
pub struct CustomMigrator {
    migrations: Vec<Box<dyn Migration<Postgres>>>,
}

impl Info<Postgres, ()> for CustomMigrator {
    fn migrations(&self) -> &Vec<Box<dyn Migration<Postgres>>> {
        &self.migrations
    }

    fn migrations_mut(&mut self) -> &mut Vec<Box<dyn Migration<Postgres>>> {
        &mut self.migrations
    }

    fn state(&self) -> &() {
        &()
    }
}

#[async_trait::async_trait]
impl DatabaseOperation<Postgres, ()> for CustomMigrator {
    async fn ensure_migration_table_exists(
        &self,
        connection: &mut <Postgres as sqlx::Database>::Connection,
    ) -> Result<(), Error> {
        sqlx::query(
            "CREATE TABLE IF NOT EXISTS _custom_table_name (
        id INT PRIMARY KEY NOT NULL GENERATED ALWAYS AS IDENTITY,
        app TEXT NOT NULL,
        name TEXT NOT NULL,
        applied_time TIMESTAMPTZ NOT NULL DEFAULT now(),
        UNIQUE (app, name)
    )",
        )
        .execute(connection)
        .await?;
        Ok(())
    }

    async fn drop_migration_table_if_exists(
        &self,
        connection: &mut <Postgres as sqlx::Database>::Connection,
    ) -> Result<(), Error> {
        sqlx::query("DROP TABLE IF EXISTS _custom_table_name")
            .execute(connection)
            .await?;
        Ok(())
    }

    async fn add_migration_to_db_table(
        &self,
        migration: &Box<dyn Migration<Postgres>>,
        connection: &mut <Postgres as sqlx::Database>::Connection,
    ) -> Result<(), Error> {
        sqlx::query("INSERT INTO _custom_table_name(app, name) VALUES ($1, $2)")
            .bind(migration.app())
            .bind(migration.name())
            .execute(connection)
            .await?;
        Ok(())
    }

    async fn delete_migration_from_db_table(
        &self,
        migration: &Box<dyn Migration<Postgres>>,
        connection: &mut <Postgres as sqlx::Database>::Connection,
    ) -> Result<(), Error> {
        sqlx::query("DELETE FROM _custom_table_name WHERE app = $1 AND name = $2")
            .bind(migration.app())
            .bind(migration.name())
            .execute(connection)
            .await?;
        Ok(())
    }

    async fn fetch_applied_migration_from_db(
        &self,
        connection: &mut <Postgres as sqlx::Database>::Connection,
    ) -> Result<Vec<AppliedMigrationSqlRow>, Error> {
        Ok(sqlx::query_as::<_, AppliedMigrationSqlRow>(
            "SELECT id, app, name, applied_time FROM _custom_table_name",
        )
        .fetch_all(connection)
        .await?)
    }

    async fn lock(
        &self,
        connection: &mut <Postgres as sqlx::Database>::Connection,
    ) -> Result<(), Error> {
        let (database_name,): (String,) = sqlx::query_as("SELECT CURRENT_DATABASE()")
            .fetch_one(&mut *connection)
            .await?;
        let lock_id = i64::from(crc32fast::hash(database_name.as_bytes()));
        sqlx::query("SELECT pg_advisory_lock($1)")
            .bind(lock_id)
            .execute(connection)
            .await?;
        Ok(())
    }

    async fn unlock(
        &self,
        connection: &mut <Postgres as sqlx::Database>::Connection,
    ) -> Result<(), Error> {
        let (database_name,): (String,) = sqlx::query_as("SELECT CURRENT_DATABASE()")
            .fetch_one(&mut *connection)
            .await?;
        let lock_id = i64::from(crc32fast::hash(database_name.as_bytes()));
        sqlx::query("SELECT pg_advisory_unlock($1)")
            .bind(lock_id)
            .execute(connection)
            .await?;
        Ok(())
    }
}
impl Migrate<Postgres, ()> for CustomMigrator {}

Structs

• Migrator
  Migrator struct which store migrations graph and information related to different library supported migrations
• Plan
  Struct which determine type of plan to use

Traits

• DatabaseOperation
  Trait which is implemented for database for performing different operations/action on database. Usually this trait is implemented for database to support database along with info trait
• Info
  Info trait which implements some of database agnostic methods to add migration or returns immutable or mutable migrations list
• Migrate
  Migrate trait which migrate a database according to requirements. This trait implements all methods which depends on DatabaseOperation trait and Info trait. This trait doesn’t requires to implement any method since all function have default implementation and all methods are database agnostics