parrot - v1.0.1

🦜 Parrot / type-safe SQL in gleam

Package Version Hex Docs erlang

Features

Most of the heavy lifting features are provided by / built into sqlc, so I do not aim to take credit for them.

✅ Supports Sqlite, PostgreSQL and MySQL.
✅ Named parameters.*1
✅ Multiple queries per file.
✅ Utility wrappers for popular gleam database libraries (lpil/sqlight, lpil/pog).
✅ Automatically pulls schema of your database.
✅ Automatically downloads sqlc binary.

*1: meaning that it infers the names of the parameters from your sql queries in the gleam function you call. f.e. WHERE username = $1 can yield sql.get_user(username:). if you have multiple parameters of the same data types this can avoid confusion and bugs.

Usage / Getting Started

Installation

$ gleam add parrot

If you target JavaScript

So here is the catch: you can only execute parrot in an erlang gleam application. However the generated code will also run in a javascript environment. So if you need parrot for a javascript project, you can create a separate package and copy over the generated module and that will work.

Define your Queries

Here are some links to help you start out, if you are unfamiliar with the sqlc annotation syntax:

Here is an example of the file structure:

├── gleam.toml
├── README.md
├── src
│   ├── app.gleam
│   └── sql
│       ├── auth.sql
│       └── posts.sql
└── test
   └── app_test.gleam

Code Generation

# automatically detects database & engine from env (DATABASE_URL by default)
$ gleam run -m parrot

# provide connection string from different environment variable
$ gleam run -m parrot -- -e PG_DATABASE_URL

# specify sqlite file
$ gleam run -m parrot -- --sqlite <file_path>

# see all options
$ gleam run -m parrot help

If you use MySQL, you also need mysqldump (comes by default if you have a mysql client installed)

If you use PostgreSQL, you also need pg_dump (comes by default if you have a postgresql client installed)

Run it!

You now have type safe access to your sql queries. You might have to write 1-2 wrapper functions for the database client library of your choice.

If you are using lpil/pog or lpil/sqlight, you are in luck! You can find functions to copy & paste into your codebase here: wrappers

An example with lpil/sqlight:

import app/sql
import parrot/dev

fn parrot_to_sqlight(param: dev.Param) -> sqlight.Value {
  // ...
}

pub fn main() {
  // ...

  let #(sql, with, expecting) = sql.get_user_by_username("alice")
  let with = parrot_to_sqlight(with)
  let row = sqlight.query(sql, on:, with:, expecting:)

  // ...
}

Examples

If you want to see how this library works in action, take a look at the integration tests:

Development

Database

There are scripts to spawn a MySQL or PostgreSQL docker container:

For example:

$ ./bin/mysql.sh
# or
$ ./bin/psql.sh

Integration Test Suite

$ just test-sqlite
$ just test-mysql
$ just test-psql

FAQ

What flavour of SQL does parrot support?

This library supports everything that sqlc supports. As the time of this writing that would be MySQL, PostgreSQL and Sqlite.

You can read more on language & SQL support here: https://docs.sqlc.dev/en/stable/reference/language-support.html

What sqlc features are not supported?

Future Work

Here are some ideas and thoughts on how you might develop this library further:

Contributions are welcomed!

Acknowledgements

✨ Search Document
OSZAR »