Home Original page

pgsql-parser/packages/deparser at main · constructive-io/pgsql-parser

pgsql-deparser is the lightning-fast, pure TypeScript solution for converting PostgreSQL ASTs back into SQL queries. Perfect companion to pgsql-parser, this focused tool delivers SQL generation without any native dependencies or WebAssembly overhead.

Installation

npm install pgsql-deparser

Features

  • ⚡ Pure TypeScript Performance – Zero runtime dependencies, no WASM, no compilation - just blazing fast SQL generation
  • 🪶 Ultra Lightweight – Minimal footprint with laser-focused functionality for AST-to-SQL conversion only
  • 🧪 Battle-Tested Reliability – Validated against 23,000+ SQL statements ensuring production-grade stability
  • 🌍 Universal Compatibility – Runs anywhere JavaScript does - browsers, Node.js, edge functions, you name it

Deparser Example

The pgsql-deparser module serializes ASTs to SQL in pure TypeScript, avoiding the full parser's native dependencies. It's useful when only SQL string conversion from ASTs is needed, and is written in pure TypeScript for easy cross-environment deployment.

Here's how you can use the deparser in your TypeScript code, using @pgsql/utils to create an AST for deparse:

import * as t from '@pgsql/utils';
import { RangeVar, SelectStmt } from '@pgsql/types';
import { deparseSync as deparse } from 'pgsql-deparser';

// This could have been obtained from any JSON or AST, not necessarily @pgsql/utils
const stmt: { SelectStmt: SelectStmt } = t.nodes.selectStmt({
  targetList: [
    t.nodes.resTarget({
      val: t.nodes.columnRef({
        fields: [t.nodes.aStar()]
      })
    })
  ],
  fromClause: [
    t.nodes.rangeVar({
      relname: 'some_table',
      inh: true,
      relpersistence: 'p'
    })
  ],
  limitOption: 'LIMIT_OPTION_DEFAULT',
  op: 'SETOP_NONE'
});

// Modify the AST if needed  
(stmt.SelectStmt.fromClause[0] as {RangeVar: RangeVar}).RangeVar.relname = 'another_table';

// Deparse the modified AST back to a SQL string
console.log(deparse(stmt));

// Output: SELECT * FROM another_table

Latest Version (PostgreSQL 17)

npm install pgsql-deparser

Version-Specific Packages (PostgreSQL 13-16)

While we highly recommend using PG17, for PostgreSQL versions 13-16, use the version-specific packages:

npm install pgsql-deparser@pg13  # PostgreSQL 13
npm install pgsql-deparser@pg14  # PostgreSQL 14
npm install pgsql-deparser@pg15  # PostgreSQL 15
npm install pgsql-deparser@pg16  # PostgreSQL 16

Version Status:

  • PG17: 🚀 Recommended (stable + modern AST)
  • PG14-16: ⚠️ Experimental (modern AST, hardening in progress)
  • PG13: Stable (legacy AST format)

Options

The deparser accepts optional configuration for formatting and output control:

import { deparseSync as deparse } from 'pgsql-deparser';

const options = {
  pretty: true,           // Enable pretty formatting (default: true)
  newline: '\n',         // Newline character (default: '\n')
  tab: '  ',             // Tab/indentation character (default: '  ')
  semicolons: true       // Add semicolons to statements (default: true)
};

const sql = deparse(ast, options);
Option Type Default Description
pretty boolean true Enable pretty formatting with indentation and line breaks
newline string '\n' Character(s) used for line breaks
tab string ' ' Character(s) used for indentation
semicolons boolean true Add semicolons to SQL statements

Pretty formatting example:

// Without pretty formatting
const sql1 = deparse(selectAst, { pretty: false });
// "SELECT id, name FROM users WHERE active = true;"

// With pretty formatting  
const sql2 = deparse(selectAst, { pretty: true });
// SELECT
//   id,
//   name
// FROM users
// WHERE
//   active = true;

For complete documentation and advanced options, see DEPARSER_USAGE.md.

Why Use pgsql-deparser?

pgsql-deparser is particularly useful in development environments where native dependencies are problematic or in applications where only the deparser functionality is required. Its independence from the full pgsql-parser package allows for more focused and lightweight SQL generation tasks.

Credits

Built on the excellent work of several contributors: