• pg.Client

new Client

new Client(config: Config)

Every field of the config object is entirely optional. A Client instance will use environment variables for all missing values.

type Config = {
  user?: string, // default process.env.PGUSER || process.env.USER
  password?: string or function, //default process.env.PGPASSWORD
  host?: string, // default process.env.PGHOST
  port?: number, // default process.env.PGPORT
  database?: string, // default process.env.PGDATABASE || user
  connectionString?: string, // e.g. postgres://user:password@host:5432/database
  ssl?: any, // passed directly to node.TLSSocket, supports all tls.connect options
  types?: any, // custom type parsers
  statement_timeout?: number, // number of milliseconds before a statement in query will time out, default is no timeout
  query_timeout?: number, // number of milliseconds before a query call will timeout, default is no timeout
  lock_timeout?: number, // number of milliseconds a query is allowed to be en lock state before it's cancelled due to lock timeout
  application_name?: string, // The name of the application that created this Client instance
  connectionTimeoutMillis?: number, // number of milliseconds to wait for connection, default is no timeout
  idle_in_transaction_session_timeout?: number // number of milliseconds before terminating any session with an open idle transaction, default is no timeout

example to create a client with specific connection information:

import pg from 'pg'
const { Client } = pg
const client = new Client({
  user: 'database-user',
  password: 'secretpassword!!',
  host: 'my.database-server.com',
  port: 5334,
  database: 'database-name',


import pg from 'pg'
const { Client } = pg
const client = new Client()
await client.connect()



You can pass an object to client.query with the signature of:

type QueryConfig {
  // the raw query text
  text: string;
  // an array of query parameters
  values?: Array<any>;
  // name of the query - used for prepared statements
  name?: string;
  // by default rows come out as a key/value pair for each row
  // pass the string 'array' here to receive rows as an array of values
  rowMode?: string;
  // custom type parsers just for this query result
  types?: Types;
  // TODO: document
  queryMode?: string;
client.query(text: string, values?: any[]) => Promise<Result>

Plain text query

import pg from 'pg'
const { Client } = pg
const client = new Client()
await client.connect()
const result = await client.query('SELECT NOW()')
await client.end()

Parameterized query

import pg from 'pg'
const { Client } = pg
const client = new Client()
await client.connect()
const result = await client.query('SELECT $1::text as name', ['brianc'])
await client.end()
client.query(config: QueryConfig) => Promise<Result>

client.query with a QueryConfig

If you pass a name parameter to the client.query method, the client will create a prepared statement.

const query = {
  name: 'get-name',
  text: 'SELECT $1::text',
  values: ['brianc'],
  rowMode: 'array',
const result = await client.query(query)
console.log(result.rows) // ['brianc']
await client.end()

client.query with a Submittable

If you pass an object to client.query and the object has a .submit function on it, the client will pass it's PostgreSQL server connection to the object and delegate query dispatching to the supplied object. This is an advanced feature mostly intended for library authors. It is incidentally also currently how the callback and promise based queries above are handled internally, but this is subject to change. It is also how pg-cursor and pg-query-stream work.

import pg from 'pg'
const { Query } = pg
const query = new Query('select $1::text as name', ['brianc'])
const result = client.query(query)
assert(query === result) // true
query.on('row', (row) => {
  console.log('row!', row) // { name: 'brianc' }
query.on('end', () => {
  console.log('query done')
query.on('error', (err) => {


Disconnects the client from the PostgreSQL server.

await client.end()
console.log('client has disconnected')



client.on('error', (err: Error) => void) => void

When the client is in the process of connecting, dispatching a query, or disconnecting it will catch and forward errors from the PostgreSQL server to the respective client.connect client.query or client.end promise; however, the client maintains a long-lived connection to the PostgreSQL back-end and due to network partitions, back-end crashes, fail-overs, etc the client can (and over a long enough time period will) eventually be disconnected while it is idle. To handle this you may want to attach an error listener to a client to catch errors. Here's a contrived example:

const client = new pg.Client()
client.on('error', (err) => {
  console.error('something bad has happened!', err.stack)
// walk over to server, unplug network cable
// process output: 'something bad has happened!' followed by stacktrace :P


client.on('end') => void

When the client disconnects from the PostgreSQL server it will emit an end event once.


Used for listen/notify events:

type Notification {
  processId: number,
  channel: string,
  payload?: string
const client = new pg.Client()
await client.connect()
client.query('LISTEN foo')
client.on('notification', (msg) => {
  console.log(msg.channel) // foo
  console.log(msg.payload) // bar!
client.query(`NOTIFY foo, 'bar!'`)


client.on('notice', (notice: Error) => void) => void

Used to log out notice messages from the PostgreSQL server.

client.on('notice', (msg) => console.warn('notice:', msg))