node-postgres 0,0,2,4,0,3,4,3 travis-ci npm

PostgreSQL client for node.js.


Build Status Dependency Status NPM version NPM downloads

Non-blocking PostgreSQL client for node.js. Pure JavaScript and optional native libpq bindings.


$ npm install pg

Intro & Examples

There are 3 ways of executing queries

  1. Passing the query to a pool
  2. Borrowing a client from a pool and executing the query with it
  3. Obtaining an exclusive client and executing the query with it

It is recommended to pass the query to a pool as often as possible. If that isn’t possible, because of long and complex transactions for example, borrow a client from a pool. Just remember to initialize the pool only once in your code so you maximize reusability of connections.

Why pooling?

If you’re working on something like a web application which makes frequent queries you’ll want to access the PostgreSQL server through a pool of clients. Why? For one thing, there is ~20-30 millisecond delay (YMMV) when connecting a new client to the PostgreSQL server because of the startup handshake. Furthermore, PostgreSQL can support only a limited number of clients…it depends on the amount of ram on your database server, but generally more than 100 clients at a time is a very bad thing. :tm: Additionally, PostgreSQL can only execute 1 query at a time per connected client, so pipelining all queries for all requests through a single, long-lived client will likely introduce a bottleneck into your application if you need high concurrency.

With that in mind we can imagine a situation where you have a web server which connects and disconnects a new client for every web request or every query (don’t do this!). If you get only 1 request at a time everything will seem to work fine, though it will be a touch slower due to the connection overhead. Once you get >100 simultaneous requests your web server will attempt to open 100 connections to the PostgreSQL backend and :boom: you’ll run out of memory on the PostgreSQL server, your database will become unresponsive, your app will seem to hang, and everything will break. Boooo!

Good news: node-postgres ships with built in client pooling. Client pooling allows your application to use a pool of already connected clients and reuse them for each request to your application. If your app needs to make more queries than there are available clients in the pool the queries will queue instead of overwhelming your database & causing a cascading failure. :thumbsup:

node-postgres uses pg-pool to manage pooling. It bundles it and exports it for convenience. If you want, you can require('pg-pool') and use it directly - it’s the same as the constructor exported at pg.Pool.

It’s highly recommended you read the documentation for pg-pool.

Here is an up & running quickly example

For more information about config.ssl check TLS (SSL) of nodejs

Pooling example

Let’s create a pool in ./lib/db.js which will be reused across the whole project

const pg = require('pg');

// create a config to configure both pooling behavior
// and client options
// note: all config is optional and the environment variables
// will be read if the config is not present
var config = {
  user: 'foo', //env var: PGUSER
  database: 'my_db', //env var: PGDATABASE
  password: 'secret', //env var: PGPASSWORD
  host: 'localhost', // Server hosting the postgres database
  port: 5432, //env var: PGPORT
  max: 10, // max number of clients in the pool
  idleTimeoutMillis: 30000, // how long a client is allowed to remain idle before being closed

//this initializes a connection pool
//it will keep idle connections open for 30 seconds
//and set a limit of maximum 10 idle clients
const pool = new pg.Pool(config);

pool.on('error', function (err, client) {
  // if an error is encountered by a client while it sits idle in the pool
  // the pool itself will emit an error event with both the error and
  // the client which emitted the original error
  // this is a rare occurrence but can happen if there is a network partition
  // between your application and the database, the database restarts, etc.
  // and so you might want to handle it and at least log it out
  console.error('idle client error', err.message, err.stack);

//export the query method for passing queries to the pool
module.exports.query = function (text, values, callback) {
  console.log('query:', text, values);
  return pool.query(text, values, callback);

// the pool also supports checking out a client for
// multiple operations, such as a transaction
module.exports.connect = function (callback) {
  return pool.connect(callback);

Now if in ./foo.js you want to pass a query to the pool

const pool = require('./lib/db');

//to run a query we just pass it to the pool
//after we're done nothing has to be taken care of
//we don't have to return any client to the pool or close a connection
pool.query('SELECT $1::int AS number', ['2'], function(err, res) {
  if(err) {
    return console.error('error running query', err);

  console.log('number:', res.rows[0].number);

Or if in ./bar.js you want borrow a client from the pool

const pool = require('./lib/db');

//ask for a client from the pool
pool.connect(function(err, client, done) {
  if(err) {
    return console.error('error fetching client from pool', err);
  //use the client for executing the query
  client.query('SELECT $1::int AS number', ['1'], function(err, result) {
    //call `done(err)` to release the client back to the pool (or destroy it if there is an error)

    if(err) {
      return console.error('error running query', err);
    //output: 1

For more examples, including how to use a connection pool with promises and async/await see the example page in the wiki.

Obtaining an exclusive client, example

var pg = require('pg');

// instantiate a new client
// the client will read connection information from
// the same environment variables used by postgres cli tools
var client = new pg.Client();

// connect to our database
client.connect(function (err) {
  if (err) throw err;

  // execute a query on our database
  client.query('SELECT $1::text as name', ['brianc'], function (err, result) {
    if (err) throw err;

    // just print the result to the console
    console.log(result.rows[0]); // outputs: { name: 'brianc' }

    // disconnect the client
    client.end(function (err) {
      if (err) throw err;

More Documentation

Native Bindings

To install the native bindings:

$ npm install pg pg-native

node-postgres contains a pure JavaScript protocol implementation which is quite fast, but you can optionally use native bindings for a 20-30% increase in parsing speed (YMMV). Both versions are adequate for production workloads. I personally use the pure JavaScript implementation because I like knowing what’s going on all the way down to the binary on the socket, and it allows for some fancier use cases which are difficult to do with libpq. :smile:

To use the native bindings, first install pg-native. Once pg-native is installed, simply replace var pg = require('pg') with var pg = require('pg').native. Make sure any exported constructors from pg are from the native instance. Example:

var pg = require('pg').native
var Pool = require('pg').Pool // bad! this is not bound to the native client
var Client = require('pg').Client // bad! this is the pure JavaScript client

var pg = require('pg').native
var Pool = pg.Pool // good! a pool bound to the native client
var Client = pg.Client // good! this client uses libpq bindings

API differences

node-postgres abstracts over the pg-native module to provide the same interface as the pure JavaScript version. Care has been taken to keep the number of api differences between the two modules to a minimum.
However, currently some differences remain, especially : * the error object in pg-native is different : notably, the information about the postgres error code is not present in field code but in the field sqlState , and the name of a few other fields is different (see, So for example, if you rely on error.code in your application, your will have to adapt your code to work with native bindings. * the notification object has a few less properties (see * column objects have less properties (see * the modules and do not work with native bindings (you will have to require ‘pg’ to use them).

Thus, it is recommended you use either the pure JavaScript or native bindings in both development and production and don’t mix & match them in the same process - it can get confusing!


  • pure JavaScript client and native libpq bindings share the same api
  • connection pooling
  • extensible js<->postgresql data-type coercion
  • supported PostgreSQL features
    • parameterized queries
    • named statements with query plan caching
    • async notifications with LISTEN/NOTIFY
    • bulk import & export with COPY TO/COPY FROM


node-postgres is by design pretty light on abstractions. These are some handy modules we’ve been using over the years to complete the picture. Entire list can be found on wiki


:heart: contributions!

If you need help getting the tests running locally or have any questions about the code when working on a patch please feel free to email me or gchat me.

I will happily accept your pull request if it: - has tests - looks reasonable - does not break backwards compatibility

Information about the testing processes is in the wiki.

Open source belongs to all of us, and we’re all invited to participate!

Troubleshooting and FAQ

The causes and solutions to common errors can be found among the Frequently Asked Questions(FAQ)


If at all possible when you open an issue please provide - version of node - version of postgres - smallest possible snippet of code to reproduce the problem

Usually I’ll pop the code into the repo as a test. Hopefully the test fails. Then I make the test pass. Then everyone’s happy!

If you need help or run into any issues getting node-postgres to work on your system please report a bug or contact me directly. I am usually available via google-talk at my github account public email address. Remember this is a labor of love, and though I try to get back to everything sometimes life takes priority, and I might take a while. It helps if you use nice code formatting in your issue, search for existing answers before posting, and come back and close out the issue if you figure out a solution. The easier you can make it for me, the quicker I’ll try and respond to you!

If you need deeper support, have application specific questions, would like to sponsor development, or want consulting around node & postgres please send me an email, I’m always happy to discuss!

I usually tweet about any important status updates or changes to node-postgres on twitter. Follow me @briancarlson to keep up to date.


Copyright © 2010-2017 Brian Carlson ([email protected])

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.


Related Repositories



PostgreSQL client for node.js. ...



Query results from node-postgres as a readable (object) stream ...



A connection pool for node-postgres ...



Production ready full stack starter kit. Featuring: Docker, Node, Postgres, Sequelize, React, Redux, GraphQL, RXjs, Webpack 3, Hot Module Reloading, Server Side Rendering and Code Splitting. ...



The postgres backend protocol implemented in pure JS for use with node.js ...

Top Contributors

brianc booo AlexanderS hoegaarden rpedela anton-kotenko lalitkapoor benesch bjornblomqvist vitaly-t sehrope gurjeet grncdr benmonty shine-on cdauth cricri badave whitelynx eugeneware adunstan benighted phated ahtih chowey jzimmek linearray kevinburke joskuijpers drdaeman


package version
buffer-writer 1.0.1
packet-reader 0.3.1
pg-connection-string 0.1.3
pg-pool 2.*
pg-types 1.*
pgpass 1.x
semver 4.3.2
dev async 0.9.0
co 4.6.0
eslint 4.2.0
eslint-config-standard 10.2.1
eslint-plugin-import 2.7.0
eslint-plugin-node 5.1.0
eslint-plugin-promise 3.5.0
eslint-plugin-standard 3.0.1
pg-copy-streams 0.3.0


-   v6.1.0 zip tar
-   v6.0.4 zip tar
-   v6.0.3 zip tar
-   v6.0.2 zip tar
-   v6.0.1 zip tar
-   v6.0.0 zip tar
-   v5.2.0 zip tar
-   v5.1.0 zip tar
-   v5.0.0 zip tar
-   v4.5.6 zip tar
-   v4.5.5 zip tar
-   v4.5.4 zip tar
-   v4.5.3 zip tar
-   v4.5.2 zip tar
-   v4.5.1 zip tar
-   v4.5.0 zip tar
-   v4.4.6 zip tar
-   v4.4.5 zip tar
-   v4.4.4 zip tar
-   v4.4.3 zip tar
-   v4.4.2 zip tar
-   v4.4.1 zip tar
-   v4.4.0 zip tar
-   v4.3.0 zip tar
-   v4.2.0 zip tar
-   v4.1.1 zip tar
-   v4.1.0 zip tar
-   v4.0.0 zip tar
-   v3.6.3 zip tar
-   v3.6.2 zip tar