sql is a set of primitives to represent SQL tables and queries. Use these primitives to:

• Auto-generate migrations: Load SQL representations in a standardized form (sql::Schema), calculate differences between schemas (sql::Migration), and generate SQL to apply the migration (sql::Migration::to_sql).
• Build SQL queries: Represent SQL queries in a data model, to create APIs for query generation. Then, generate the SQL query. Note: this library does not support parsing SQL queries (yet).

For auto-generating migrations, there are a few built-in schema sources:

• Postgres: sql_sqlx
• OpenAPI v3: sql_openapi

If you need another source, you should define a way to build a sql::Schema from your data source, then use sql to auto-generate migrations.

Current tools that support this:

• ormlite

If you use this library, submit a PR to be added to this list.

This example reads the schema from a postgres database, defines an empty schema (which should be filled in), and then computes the migration to apply to the database.

use sql_sqlx::FromPostgres;

#[tokio::main]
async fn main() {
    let url = std::env::var("DATABASE_URL").unwrap();
    let mut conn = sqlx::postgres::PgConnection::connect(&url).await?;
    let current = Schema::try_from_postgres(&mut conn, schema_name).await?;
    let end_state = Schema::default(); // Load your end-state by manually defining it, or building it from another source
    let migration = current.migrate_to(end_state, &sql::Options::default());
    
    for statement in migration.statements {
        let statement = statement.to_sql(Dialect::Postgres);
        println!("{}", statement);
    }
}

• When calculating migrations, create commented out lines for column deletion
• ? When calculating migrations, do alter column by calculating word distance between column names