A Prettier plugin for SQL that uses sql-parser-cst and the actual Prettier formatting algorithm.

Try it live!

Like Prettier for JavaScript, this plugin formats SQL expressions differently depending on their length. A short SQL query will be formatted on a single line:

SELECT a, b, c FROM tbl WHERE x > 10;

A longer query, will get each clause printed on a separate line:

SELECT id, client.name, client.priority
FROM client
WHERE client.id IN (12, 18, 121);

An even longer one gets the contents of each clause indented:

SELECT
  client.id,
  client.name AS client_name,
  organization.name AS org_name,
  count(orders.id) AS nr_of_orders
FROM
  client
  LEFT JOIN organization ON client.organization_id = organization.id
  LEFT JOIN orders ON orders.client_id = client.id
WHERE
  client.status = 'active'
  AND client.id IN (28, 214, 457)
  AND orders.status IN ('active', 'pending', 'processing')
GROUP BY client.id
ORDER BY client.name
LIMIT 100;

• Adapt formatting based on expression length.
• Stick to one style and avoid configuration options.
• Format embedded languages (like JSON data and JavaScript programs).
• When unsure, preserve existing syntax.

Currently this plugin preserves most of the syntax elements and concentrates mainly on the layout of whitespace.

See STYLE_GUIDE for overview of the SQL formatting style used.

Install it as any other Prettier plugin:

npm install --save-dev prettier prettier-plugin-sql-cst

Then use it on SQL files through Prettier command line tool or Prettier extension for your editor of choice.

By default the plugin will determine SQL dialect based on file extension:

• .sql or .sqlite - SQLite
• .bigquery - BigQuery

You can override this behavior with a prettier configuration:

{
  "overrides": [
    {
      "files": ["*.sql"],
      "options": { "parser": "bigquery" }
    }
  ]
}

The plugin provides the following parsers:

• sqlite
• bigquery

The standard Prettier options printWidth, tabWidth, useTabs apply. There are also some SQL-specific options:

Currently this plugin supports two SQL dialects:

• SQLite - full support.
• BigQuery - full support.

The specifics of the SQL formatting style are still very much subject to change. Though the general principles should be mostly in place by now.