/
opt
/
alt
/
alt-nodejs10
/
root
/
lib
/
node_modules
/
npm
/
html
/
html
/
api
/
Upload Filee
HOME
<!doctype html> <html lang="en"> <head> <meta charset="utf-8"> <meta name="viewport" content="width=device-width"> <title>Stream | Node.js v10.24.1 Documentation</title> <link rel="stylesheet" href="https://fonts.googleapis.com/css?family=Lato:400,700,400italic"> <link rel="stylesheet" href="assets/style.css"> <link rel="stylesheet" href="assets/sh.css"> <link rel="canonical" href="https://nodejs.org/api/stream.html"> </head> <body class="alt apidoc" id="api-section-stream"> <div id="content" class="clearfix"> <div id="column2" class="interior"> <div id="intro" class="interior"> <a href="/" title="Go back to the home page"> Node.js </a> </div> <ul> <li><a href="documentation.html" class="nav-documentation">About these Docs</a></li> <li><a href="synopsis.html" class="nav-synopsis">Usage & Example</a></li> </ul> <div class="line"></div> <ul> <li><a href="assert.html" class="nav-assert">Assertion Testing</a></li> <li><a href="async_hooks.html" class="nav-async_hooks">Async Hooks</a></li> <li><a href="buffer.html" class="nav-buffer">Buffer</a></li> <li><a href="addons.html" class="nav-addons">C++ Addons</a></li> <li><a href="n-api.html" class="nav-n-api">C/C++ Addons - N-API</a></li> <li><a href="child_process.html" class="nav-child_process">Child Processes</a></li> <li><a href="cluster.html" class="nav-cluster">Cluster</a></li> <li><a href="cli.html" class="nav-cli">Command Line Options</a></li> <li><a href="console.html" class="nav-console">Console</a></li> <li><a href="crypto.html" class="nav-crypto">Crypto</a></li> <li><a href="debugger.html" class="nav-debugger">Debugger</a></li> <li><a href="deprecations.html" class="nav-deprecations">Deprecated APIs</a></li> <li><a href="dns.html" class="nav-dns">DNS</a></li> <li><a href="domain.html" class="nav-domain">Domain</a></li> <li><a href="esm.html" class="nav-esm">ECMAScript Modules</a></li> <li><a href="errors.html" class="nav-errors">Errors</a></li> <li><a href="events.html" class="nav-events">Events</a></li> <li><a href="fs.html" class="nav-fs">File System</a></li> <li><a href="globals.html" class="nav-globals">Globals</a></li> <li><a href="http.html" class="nav-http">HTTP</a></li> <li><a href="http2.html" class="nav-http2">HTTP/2</a></li> <li><a href="https.html" class="nav-https">HTTPS</a></li> <li><a href="inspector.html" class="nav-inspector">Inspector</a></li> <li><a href="intl.html" class="nav-intl">Internationalization</a></li> <li><a href="modules.html" class="nav-modules">Modules</a></li> <li><a href="net.html" class="nav-net">Net</a></li> <li><a href="os.html" class="nav-os">OS</a></li> <li><a href="path.html" class="nav-path">Path</a></li> <li><a href="perf_hooks.html" class="nav-perf_hooks">Performance Hooks</a></li> <li><a href="process.html" class="nav-process">Process</a></li> <li><a href="punycode.html" class="nav-punycode">Punycode</a></li> <li><a href="querystring.html" class="nav-querystring">Query Strings</a></li> <li><a href="readline.html" class="nav-readline">Readline</a></li> <li><a href="repl.html" class="nav-repl">REPL</a></li> <li><a href="stream.html" class="nav-stream active">Stream</a></li> <li><a href="string_decoder.html" class="nav-string_decoder">String Decoder</a></li> <li><a href="timers.html" class="nav-timers">Timers</a></li> <li><a href="tls.html" class="nav-tls">TLS/SSL</a></li> <li><a href="tracing.html" class="nav-tracing">Trace Events</a></li> <li><a href="tty.html" class="nav-tty">TTY</a></li> <li><a href="dgram.html" class="nav-dgram">UDP/Datagram</a></li> <li><a href="url.html" class="nav-url">URL</a></li> <li><a href="util.html" class="nav-util">Utilities</a></li> <li><a href="v8.html" class="nav-v8">V8</a></li> <li><a href="vm.html" class="nav-vm">VM</a></li> <li><a href="worker_threads.html" class="nav-worker_threads">Worker Threads</a></li> <li><a href="zlib.html" class="nav-zlib">Zlib</a></li> </ul> <div class="line"></div> <ul> <li><a href="https://github.com/nodejs/node" class="nav-https-github-com-nodejs-node">GitHub Repo & Issue Tracker</a></li> </ul> </div> <div id="column1" data-id="stream" class="interior"> <header> <h1>Node.js v10.24.1 Documentation</h1> <div id="gtoc"> <ul> <li> <a href="index.html" name="toc">Index</a> </li> <li> <a href="all.html">View on single page</a> </li> <li> <a href="stream.json">View as JSON</a> </li> <li class="version-picker"> <a href="#">View another version <span>▼</span></a> <ol class="version-picker"><li><a href="https://nodejs.org/docs/latest-v15.x/api/stream.html">15.x</a></li> <li><a href="https://nodejs.org/docs/latest-v14.x/api/stream.html">14.x <b>LTS</b></a></li> <li><a href="https://nodejs.org/docs/latest-v13.x/api/stream.html">13.x</a></li> <li><a href="https://nodejs.org/docs/latest-v12.x/api/stream.html">12.x <b>LTS</b></a></li> <li><a href="https://nodejs.org/docs/latest-v11.x/api/stream.html">11.x</a></li> <li><a href="https://nodejs.org/docs/latest-v10.x/api/stream.html">10.x <b>LTS</b></a></li> <li><a href="https://nodejs.org/docs/latest-v9.x/api/stream.html">9.x</a></li> <li><a href="https://nodejs.org/docs/latest-v8.x/api/stream.html">8.x</a></li> <li><a href="https://nodejs.org/docs/latest-v7.x/api/stream.html">7.x</a></li> <li><a href="https://nodejs.org/docs/latest-v6.x/api/stream.html">6.x</a></li> <li><a href="https://nodejs.org/docs/latest-v5.x/api/stream.html">5.x</a></li> <li><a href="https://nodejs.org/docs/latest-v4.x/api/stream.html">4.x</a></li> <li><a href="https://nodejs.org/docs/latest-v0.12.x/api/stream.html">0.12.x</a></li> <li><a href="https://nodejs.org/docs/latest-v0.10.x/api/stream.html">0.10.x</a></li></ol> </li> <li class="edit_on_github"><a href="https://github.com/nodejs/node/edit/master/doc/api/stream.md"><span class="github_icon"><svg height="16" width="16" viewBox="0 0 16.1 16.1" fill="currentColor"><path d="M8 0a8 8 0 0 0-2.5 15.6c.4 0 .5-.2.5-.4v-1.5c-2 .4-2.5-.5-2.7-1 0-.1-.5-.9-.8-1-.3-.2-.7-.6 0-.6.6 0 1 .6 1.2.8.7 1.2 1.9 1 2.4.7 0-.5.2-.9.5-1-1.8-.3-3.7-1-3.7-4 0-.9.3-1.6.8-2.2 0-.2-.3-1 .1-2 0 0 .7-.3 2.2.7a7.4 7.4 0 0 1 4 0c1.5-1 2.2-.8 2.2-.8.5 1.1.2 2 .1 2.1.5.6.8 1.3.8 2.2 0 3-1.9 3.7-3.6 4 .3.2.5.7.5 1.4v2.2c0 .2.1.5.5.4A8 8 0 0 0 16 8a8 8 0 0 0-8-8z"/></svg></span>Edit on GitHub</a></li> </ul> </div> <hr> </header> <div id="toc"> <h2>Table of Contents</h2> <ul> <li> <p><span class="stability_2"><a href="#stream_stream">Stream</a></span></p> <ul> <li><a href="#stream_organization_of_this_document">Organization of this Document</a></li> <li> <p><a href="#stream_types_of_streams">Types of Streams</a></p> <ul> <li><a href="#stream_object_mode">Object Mode</a></li> <li><a href="#stream_buffering">Buffering</a></li> </ul> </li> <li> <p><a href="#stream_api_for_stream_consumers">API for Stream Consumers</a></p> <ul> <li> <p><a href="#stream_writable_streams">Writable Streams</a></p> <ul> <li> <p><a href="#stream_class_stream_writable">Class: stream.Writable</a></p> <ul> <li><a href="#stream_event_close">Event: 'close'</a></li> <li><a href="#stream_event_drain">Event: 'drain'</a></li> <li><a href="#stream_event_error">Event: 'error'</a></li> <li><a href="#stream_event_finish">Event: 'finish'</a></li> <li><a href="#stream_event_pipe">Event: 'pipe'</a></li> <li><a href="#stream_event_unpipe">Event: 'unpipe'</a></li> <li><a href="#stream_writable_cork">writable.cork()</a></li> <li><a href="#stream_writable_destroy_error">writable.destroy([error])</a></li> <li><a href="#stream_writable_end_chunk_encoding_callback">writable.end([chunk][, encoding][, callback])</a></li> <li><a href="#stream_writable_setdefaultencoding_encoding">writable.setDefaultEncoding(encoding)</a></li> <li><a href="#stream_writable_uncork">writable.uncork()</a></li> <li><a href="#stream_writable_writable">writable.writable</a></li> <li><a href="#stream_writable_writablehighwatermark">writable.writableHighWaterMark</a></li> <li><a href="#stream_writable_writablelength">writable.writableLength</a></li> <li><a href="#stream_writable_write_chunk_encoding_callback">writable.write(chunk[, encoding][, callback])</a></li> </ul> </li> </ul> </li> <li> <p><a href="#stream_readable_streams">Readable Streams</a></p> <ul> <li><a href="#stream_two_reading_modes">Two Reading Modes</a></li> <li><a href="#stream_three_states">Three States</a></li> <li><a href="#stream_choose_one_api_style">Choose One API Style</a></li> <li> <p><a href="#stream_class_stream_readable">Class: stream.Readable</a></p> <ul> <li><a href="#stream_event_close_1">Event: 'close'</a></li> <li><a href="#stream_event_data">Event: 'data'</a></li> <li><a href="#stream_event_end">Event: 'end'</a></li> <li><a href="#stream_event_error_1">Event: 'error'</a></li> <li><a href="#stream_event_readable">Event: 'readable'</a></li> <li><a href="#stream_readable_destroy_error">readable.destroy([error])</a></li> <li><a href="#stream_readable_ispaused">readable.isPaused()</a></li> <li><a href="#stream_readable_pause">readable.pause()</a></li> <li><a href="#stream_readable_pipe_destination_options">readable.pipe(destination[, options])</a></li> <li><a href="#stream_readable_read_size">readable.read([size])</a></li> <li><a href="#stream_readable_readable">readable.readable</a></li> <li><a href="#stream_readable_readableflowing">readable.readableFlowing</a></li> <li><a href="#stream_readable_readablehighwatermark">readable.readableHighWaterMark</a></li> <li><a href="#stream_readable_readablelength">readable.readableLength</a></li> <li><a href="#stream_readable_resume">readable.resume()</a></li> <li><a href="#stream_readable_setencoding_encoding">readable.setEncoding(encoding)</a></li> <li><a href="#stream_readable_unpipe_destination">readable.unpipe([destination])</a></li> <li><a href="#stream_readable_unshift_chunk">readable.unshift(chunk)</a></li> <li><a href="#stream_readable_wrap_stream">readable.wrap(stream)</a></li> <li><span class="stability_2"><a href="#stream_readable_symbol_asynciterator">readable[Symbol.asyncIterator]()</a></span></li> </ul> </li> </ul> </li> <li> <p><a href="#stream_duplex_and_transform_streams">Duplex and Transform Streams</a></p> <ul> <li><a href="#stream_class_stream_duplex">Class: stream.Duplex</a></li> <li> <p><a href="#stream_class_stream_transform">Class: stream.Transform</a></p> <ul> <li><a href="#stream_transform_destroy_error">transform.destroy([error])</a></li> </ul> </li> </ul> </li> <li><a href="#stream_stream_finished_stream_callback">stream.finished(stream, callback)</a></li> <li><a href="#stream_stream_pipeline_streams_callback">stream.pipeline(...streams[, callback])</a></li> <li><a href="#stream_readable_from_iterable_options">Readable.from(iterable, [options])</a></li> </ul> </li> <li> <p><a href="#stream_api_for_stream_implementers">API for Stream Implementers</a></p> <ul> <li><a href="#stream_simplified_construction">Simplified Construction</a></li> <li> <p><a href="#stream_implementing_a_writable_stream">Implementing a Writable Stream</a></p> <ul> <li><a href="#stream_constructor_new_stream_writable_options">Constructor: new stream.Writable([options])</a></li> <li><a href="#stream_writable_write_chunk_encoding_callback_1">writable._write(chunk, encoding, callback)</a></li> <li><a href="#stream_writable_writev_chunks_callback">writable._writev(chunks, callback)</a></li> <li><a href="#stream_writable_destroy_err_callback">writable._destroy(err, callback)</a></li> <li><a href="#stream_writable_final_callback">writable._final(callback)</a></li> <li><a href="#stream_errors_while_writing">Errors While Writing</a></li> <li><a href="#stream_an_example_writable_stream">An Example Writable Stream</a></li> <li><a href="#stream_decoding_buffers_in_a_writable_stream">Decoding buffers in a Writable Stream</a></li> </ul> </li> <li> <p><a href="#stream_implementing_a_readable_stream">Implementing a Readable Stream</a></p> <ul> <li><a href="#stream_new_stream_readable_options">new stream.Readable([options])</a></li> <li><a href="#stream_readable_read_size_1">readable._read(size)</a></li> <li><a href="#stream_readable_destroy_err_callback">readable._destroy(err, callback)</a></li> <li><a href="#stream_readable_push_chunk_encoding">readable.push(chunk[, encoding])</a></li> <li><a href="#stream_errors_while_reading">Errors While Reading</a></li> <li><a href="#stream_an_example_counting_stream">An Example Counting Stream</a></li> </ul> </li> <li> <p><a href="#stream_implementing_a_duplex_stream">Implementing a Duplex Stream</a></p> <ul> <li><a href="#stream_new_stream_duplex_options">new stream.Duplex(options)</a></li> <li><a href="#stream_an_example_duplex_stream">An Example Duplex Stream</a></li> <li><a href="#stream_object_mode_duplex_streams">Object Mode Duplex Streams</a></li> </ul> </li> <li> <p><a href="#stream_implementing_a_transform_stream">Implementing a Transform Stream</a></p> <ul> <li><a href="#stream_new_stream_transform_options">new stream.Transform([options])</a></li> <li><a href="#stream_events_finish_and_end">Events: 'finish' and 'end'</a></li> <li><a href="#stream_transform_flush_callback">transform._flush(callback)</a></li> <li><a href="#stream_transform_transform_chunk_encoding_callback">transform._transform(chunk, encoding, callback)</a></li> <li><a href="#stream_class_stream_passthrough">Class: stream.PassThrough</a></li> </ul> </li> </ul> </li> <li> <p><a href="#stream_additional_notes">Additional Notes</a></p> <ul> <li> <p><a href="#stream_streams_compatibility_with_async_generators_and_async_iterators">Streams Compatibility with Async Generators and Async Iterators</a></p> <ul> <li><a href="#stream_consuming_readable_streams_with_async_iterators">Consuming Readable Streams with Async Iterators</a></li> <li><a href="#stream_creating_readable_streams_with_async_generators">Creating Readable Streams with Async Generators</a></li> <li><a href="#stream_piping_to_writable_streams_from_async_iterators">Piping to Writable Streams from Async Iterators</a></li> </ul> </li> <li><a href="#stream_compatibility_with_older_node_js_versions">Compatibility with Older Node.js Versions</a></li> <li><a href="#stream_readable_read_0"><code>readable.read(0)</code></a></li> <li><a href="#stream_readable_push"><code>readable.push('')</code></a></li> <li><a href="#stream_highwatermark_discrepancy_after_calling_readable_setencoding"><code>highWaterMark</code> discrepancy after calling <code>readable.setEncoding()</code></a></li> </ul> </li> </ul> </li> </ul> </div> <div id="apicontent"> <h1>Stream<a class="srclink" href="https://github.com/nodejs/node/blob/5182a7ece0b71feeb9157f7aa348a15d53e32058/lib/stream.js#L30">[src]</a><span><a class="mark" href="#stream_stream" id="stream_stream">#</a></span></h1> <p></p><div class="api_stability api_stability_2"><a href="documentation.html#documentation_stability_index">Stability: 2</a> - Stable</div><p></p> <p>A stream is an abstract interface for working with streaming data in Node.js. The <code>stream</code> module provides a base API that makes it easy to build objects that implement the stream interface.</p> <p>There are many stream objects provided by Node.js. For instance, a <a href="http.html#http_class_http_incomingmessage">request to an HTTP server</a> and <a href="process.html#process_process_stdout"><code>process.stdout</code></a> are both stream instances.</p> <p>Streams can be readable, writable, or both. All streams are instances of <a href="events.html#events_class_eventemitter"><code>EventEmitter</code></a>.</p> <p>The <code>stream</code> module can be accessed using:</p> <pre><code class="language-js">const stream = require('stream'); </code></pre> <p>While it is important to understand how streams work, the <code>stream</code> module itself is most useful for developers that are creating new types of stream instances. Developers who are primarily <em>consuming</em> stream objects will rarely need to use the <code>stream</code> module directly.</p> <h2>Organization of this Document<span><a class="mark" href="#stream_organization_of_this_document" id="stream_organization_of_this_document">#</a></span></h2> <p>This document is divided into two primary sections with a third section for additional notes. The first section explains the elements of the stream API that are required to <em>use</em> streams within an application. The second section explains the elements of the API that are required to <em>implement</em> new types of streams.</p> <h2>Types of Streams<span><a class="mark" href="#stream_types_of_streams" id="stream_types_of_streams">#</a></span></h2> <p>There are four fundamental stream types within Node.js:</p> <ul> <li><a href="#stream_class_stream_writable"><code>Writable</code></a> - streams to which data can be written (for example, <a href="fs.html#fs_fs_createwritestream_path_options"><code>fs.createWriteStream()</code></a>).</li> <li><a href="#stream_class_stream_readable"><code>Readable</code></a> - streams from which data can be read (for example, <a href="fs.html#fs_fs_createreadstream_path_options"><code>fs.createReadStream()</code></a>).</li> <li><a href="#stream_class_stream_duplex"><code>Duplex</code></a> - streams that are both <code>Readable</code> and <code>Writable</code> (for example, <a href="net.html#net_class_net_socket"><code>net.Socket</code></a>).</li> <li><a href="#stream_class_stream_transform"><code>Transform</code></a> - <code>Duplex</code> streams that can modify or transform the data as it is written and read (for example, <a href="zlib.html#zlib_zlib_createdeflate_options"><code>zlib.createDeflate()</code></a>).</li> </ul> <p>Additionally, this module includes the utility functions <a href="#stream_stream_pipeline_streams_callback">pipeline</a>, <a href="#stream_stream_finished_stream_callback">finished</a> and <a href="#readable.from">Readable.from</a>.</p> <h3>Object Mode<span><a class="mark" href="#stream_object_mode" id="stream_object_mode">#</a></span></h3> <p>All streams created by Node.js APIs operate exclusively on strings and <code>Buffer</code> (or <code>Uint8Array</code>) objects. It is possible, however, for stream implementations to work with other types of JavaScript values (with the exception of <code>null</code>, which serves a special purpose within streams). Such streams are considered to operate in "object mode".</p> <p>Stream instances are switched into object mode using the <code>objectMode</code> option when the stream is created. Attempting to switch an existing stream into object mode is not safe.</p> <h3>Buffering<span><a class="mark" href="#stream_buffering" id="stream_buffering">#</a></span></h3> <p>Both <a href="#stream_class_stream_writable"><code>Writable</code></a> and <a href="#stream_class_stream_readable"><code>Readable</code></a> streams will store data in an internal buffer that can be retrieved using <code>writable.writableBuffer</code> or <code>readable.readableBuffer</code>, respectively.</p> <p>The amount of data potentially buffered depends on the <code>highWaterMark</code> option passed into the stream's constructor. For normal streams, the <code>highWaterMark</code> option specifies a <a href="#stream_highwatermark_discrepancy_after_calling_readable_setencoding">total number of bytes</a>. For streams operating in object mode, the <code>highWaterMark</code> specifies a total number of objects.</p> <p>Data is buffered in <code>Readable</code> streams when the implementation calls <a href="#stream_readable_push_chunk_encoding"><code>stream.push(chunk)</code></a>. If the consumer of the Stream does not call <a href="#stream_readable_read_size"><code>stream.read()</code></a>, the data will sit in the internal queue until it is consumed.</p> <p>Once the total size of the internal read buffer reaches the threshold specified by <code>highWaterMark</code>, the stream will temporarily stop reading data from the underlying resource until the data currently buffered can be consumed (that is, the stream will stop calling the internal <code>readable._read()</code> method that is used to fill the read buffer).</p> <p>Data is buffered in <code>Writable</code> streams when the <a href="#stream_writable_write_chunk_encoding_callback"><code>writable.write(chunk)</code></a> method is called repeatedly. While the total size of the internal write buffer is below the threshold set by <code>highWaterMark</code>, calls to <code>writable.write()</code> will return <code>true</code>. Once the size of the internal buffer reaches or exceeds the <code>highWaterMark</code>, <code>false</code> will be returned.</p> <p>A key goal of the <code>stream</code> API, particularly the <a href="#stream_readable_pipe_destination_options"><code>stream.pipe()</code></a> method, is to limit the buffering of data to acceptable levels such that sources and destinations of differing speeds will not overwhelm the available memory.</p> <p>Because <a href="#stream_class_stream_duplex"><code>Duplex</code></a> and <a href="#stream_class_stream_transform"><code>Transform</code></a> streams are both <code>Readable</code> and <code>Writable</code>, each maintains <em>two</em> separate internal buffers used for reading and writing, allowing each side to operate independently of the other while maintaining an appropriate and efficient flow of data. For example, <a href="net.html#net_class_net_socket"><code>net.Socket</code></a> instances are <a href="#stream_class_stream_duplex"><code>Duplex</code></a> streams whose <code>Readable</code> side allows consumption of data received <em>from</em> the socket and whose <code>Writable</code> side allows writing data <em>to</em> the socket. Because data may be written to the socket at a faster or slower rate than data is received, it is important for each side to operate (and buffer) independently of the other.</p> <h2>API for Stream Consumers<span><a class="mark" href="#stream_api_for_stream_consumers" id="stream_api_for_stream_consumers">#</a></span></h2> <p>Almost all Node.js applications, no matter how simple, use streams in some manner. The following is an example of using streams in a Node.js application that implements an HTTP server:</p> <pre><code class="language-js">const http = require('http'); const server = http.createServer((req, res) => { // req is an http.IncomingMessage, which is a Readable Stream // res is an http.ServerResponse, which is a Writable Stream let body = ''; // Get the data as utf8 strings. // If an encoding is not set, Buffer objects will be received. req.setEncoding('utf8'); // Readable streams emit 'data' events once a listener is added req.on('data', (chunk) => { body += chunk; }); // the 'end' event indicates that the entire body has been received req.on('end', () => { try { const data = JSON.parse(body); // write back something interesting to the user: res.write(typeof data); res.end(); } catch (er) { // uh oh! bad json! res.statusCode = 400; return res.end(`error: ${er.message}`); } }); }); server.listen(1337); // $ curl localhost:1337 -d "{}" // object // $ curl localhost:1337 -d "\"foo\"" // string // $ curl localhost:1337 -d "not json" // error: Unexpected token o in JSON at position 1 </code></pre> <p><a href="#stream_class_stream_writable"><code>Writable</code></a> streams (such as <code>res</code> in the example) expose methods such as <code>write()</code> and <code>end()</code> that are used to write data onto the stream.</p> <p><a href="#stream_class_stream_readable"><code>Readable</code></a> streams use the <a href="events.html#events_class_eventemitter"><code>EventEmitter</code></a> API for notifying application code when data is available to be read off the stream. That available data can be read from the stream in multiple ways.</p> <p>Both <a href="#stream_class_stream_writable"><code>Writable</code></a> and <a href="#stream_class_stream_readable"><code>Readable</code></a> streams use the <a href="events.html#events_class_eventemitter"><code>EventEmitter</code></a> API in various ways to communicate the current state of the stream.</p> <p><a href="#stream_class_stream_duplex"><code>Duplex</code></a> and <a href="#stream_class_stream_transform"><code>Transform</code></a> streams are both <a href="#stream_class_stream_writable"><code>Writable</code></a> and <a href="#stream_class_stream_readable"><code>Readable</code></a>.</p> <p>Applications that are either writing data to or consuming data from a stream are not required to implement the stream interfaces directly and will generally have no reason to call <code>require('stream')</code>.</p> <p>Developers wishing to implement new types of streams should refer to the section <a href="#stream_api_for_stream_implementers">API for Stream Implementers</a>.</p> <h3>Writable Streams<span><a class="mark" href="#stream_writable_streams" id="stream_writable_streams">#</a></span></h3> <p>Writable streams are an abstraction for a <em>destination</em> to which data is written.</p> <p>Examples of <a href="#stream_class_stream_writable"><code>Writable</code></a> streams include:</p> <ul> <li><a href="http.html#http_class_http_clientrequest">HTTP requests, on the client</a></li> <li><a href="http.html#http_class_http_serverresponse">HTTP responses, on the server</a></li> <li><a href="fs.html#fs_class_fs_writestream">fs write streams</a></li> <li><a href="zlib.html">zlib streams</a></li> <li><a href="crypto.html">crypto streams</a></li> <li><a href="net.html#net_class_net_socket">TCP sockets</a></li> <li><a href="child_process.html#child_process_subprocess_stdin">child process stdin</a></li> <li><a href="process.html#process_process_stdout"><code>process.stdout</code></a>, <a href="process.html#process_process_stderr"><code>process.stderr</code></a></li> </ul> <p>Some of these examples are actually <a href="#stream_class_stream_duplex"><code>Duplex</code></a> streams that implement the <a href="#stream_class_stream_writable"><code>Writable</code></a> interface.</p> <p>All <a href="#stream_class_stream_writable"><code>Writable</code></a> streams implement the interface defined by the <code>stream.Writable</code> class.</p> <p>While specific instances of <a href="#stream_class_stream_writable"><code>Writable</code></a> streams may differ in various ways, all <code>Writable</code> streams follow the same fundamental usage pattern as illustrated in the example below:</p> <pre><code class="language-js">const myStream = getWritableStreamSomehow(); myStream.write('some data'); myStream.write('some more data'); myStream.end('done writing data'); </code></pre> <h4>Class: stream.Writable<span><a class="mark" href="#stream_class_stream_writable" id="stream_class_stream_writable">#</a></span></h4> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <h5>Event: 'close'<span><a class="mark" href="#stream_event_close" id="stream_event_close">#</a></span></h5> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v10.0.0</td> <td><p>Add <code>emitClose</code> option to specify if <code>'close'</code> is emitted on destroy.</p></td></tr> <tr><td>v0.9.4</td> <td><p><span>Added in: v0.9.4</span></p></td></tr> </tbody></table> </details> </div> <p>The <code>'close'</code> event is emitted when the stream and any of its underlying resources (a file descriptor, for example) have been closed. The event indicates that no more events will be emitted, and no further computation will occur.</p> <p>A <a href="#stream_class_stream_writable"><code>Writable</code></a> stream will always emit the <code>'close'</code> event if it is created with the <code>emitClose</code> option.</p> <h5>Event: 'drain'<span><a class="mark" href="#stream_event_drain" id="stream_event_drain">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <p>If a call to <a href="#stream_writable_write_chunk_encoding_callback"><code>stream.write(chunk)</code></a> returns <code>false</code>, the <code>'drain'</code> event will be emitted when it is appropriate to resume writing data to the stream.</p> <pre><code class="language-js">// Write the data to the supplied writable stream one million times. // Be attentive to back-pressure. function writeOneMillionTimes(writer, data, encoding, callback) { let i = 1000000; write(); function write() { let ok = true; do { i--; if (i === 0) { // last time! writer.write(data, encoding, callback); } else { // see if we should continue, or wait // don't pass the callback, because we're not done yet. ok = writer.write(data, encoding); } } while (i > 0 && ok); if (i > 0) { // had to stop early! // write some more once it drains writer.once('drain', write); } } } </code></pre> <h5>Event: 'error'<span><a class="mark" href="#stream_event_error" id="stream_event_error">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <ul> <li><a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Error" class="type"><Error></a></li> </ul> <p>The <code>'error'</code> event is emitted if an error occurred while writing or piping data. The listener callback is passed a single <code>Error</code> argument when called.</p> <p>The stream is not closed when the <code>'error'</code> event is emitted.</p> <h5>Event: 'finish'<span><a class="mark" href="#stream_event_finish" id="stream_event_finish">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <p>The <code>'finish'</code> event is emitted after the <a href="#stream_writable_end_chunk_encoding_callback"><code>stream.end()</code></a> method has been called, and all data has been flushed to the underlying system.</p> <pre><code class="language-js">const writer = getWritableStreamSomehow(); for (let i = 0; i < 100; i++) { writer.write(`hello, #${i}!\n`); } writer.end('This is the end\n'); writer.on('finish', () => { console.log('All writes are now complete.'); }); </code></pre> <h5>Event: 'pipe'<span><a class="mark" href="#stream_event_pipe" id="stream_event_pipe">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <ul> <li><code>src</code> <a href="stream.html#stream_class_stream_readable" class="type"><stream.Readable></a> source stream that is piping to this writable</li> </ul> <p>The <code>'pipe'</code> event is emitted when the <a href="#stream_readable_pipe_destination_options"><code>stream.pipe()</code></a> method is called on a readable stream, adding this writable to its set of destinations.</p> <pre><code class="language-js">const writer = getWritableStreamSomehow(); const reader = getReadableStreamSomehow(); writer.on('pipe', (src) => { console.log('Something is piping into the writer.'); assert.equal(src, reader); }); reader.pipe(writer); </code></pre> <h5>Event: 'unpipe'<span><a class="mark" href="#stream_event_unpipe" id="stream_event_unpipe">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <ul> <li><code>src</code> <a href="stream.html#stream_class_stream_readable" class="type"><stream.Readable></a> The source stream that <a href="#stream_readable_unpipe_destination">unpiped</a> this writable</li> </ul> <p>The <code>'unpipe'</code> event is emitted when the <a href="#stream_readable_unpipe_destination"><code>stream.unpipe()</code></a> method is called on a <a href="#stream_class_stream_readable"><code>Readable</code></a> stream, removing this <a href="#stream_class_stream_writable"><code>Writable</code></a> from its set of destinations.</p> <p>This is also emitted in case this <a href="#stream_class_stream_writable"><code>Writable</code></a> stream emits an error when a <a href="#stream_class_stream_readable"><code>Readable</code></a> stream pipes into it.</p> <pre><code class="language-js">const writer = getWritableStreamSomehow(); const reader = getReadableStreamSomehow(); writer.on('unpipe', (src) => { console.log('Something has stopped piping into the writer.'); assert.equal(src, reader); }); reader.pipe(writer); reader.unpipe(writer); </code></pre> <h5>writable.cork()<span><a class="mark" href="#stream_writable_cork" id="stream_writable_cork">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.11.2</span> </div> <p>The <code>writable.cork()</code> method forces all written data to be buffered in memory. The buffered data will be flushed when either the <a href="#stream_writable_uncork"><code>stream.uncork()</code></a> or <a href="#stream_writable_end_chunk_encoding_callback"><code>stream.end()</code></a> methods are called.</p> <p>The primary intent of <code>writable.cork()</code> is to avoid a situation where writing many small chunks of data to a stream do not cause a backup in the internal buffer that would have an adverse impact on performance. In such situations, implementations that implement the <code>writable._writev()</code> method can perform buffered writes in a more optimized manner.</p> <p>See also: <a href="#stream_writable_uncork"><code>writable.uncork()</code></a>.</p> <h5>writable.destroy([error])<span><a class="mark" href="#stream_writable_destroy_error" id="stream_writable_destroy_error">#</a></span></h5> <div class="api_metadata"> <span>Added in: v8.0.0</span> </div> <ul> <li><code>error</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Error" class="type"><Error></a></li> <li>Returns: <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/this" class="type"><this></a></li> </ul> <p>Destroy the stream, and emit the passed <code>'error'</code> and a <code>'close'</code> event. After this call, the writable stream has ended and subsequent calls to <code>write()</code> or <code>end()</code> will result in an <code>ERR_STREAM_DESTROYED</code> error. Implementors should not override this method, but instead implement <a href="#stream_writable_destroy_err_callback"><code>writable._destroy()</code></a>.</p> <h5>writable.end([chunk][, encoding][, callback])<span><a class="mark" href="#stream_writable_end_chunk_encoding_callback" id="stream_writable_end_chunk_encoding_callback">#</a></span></h5> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v10.0.0</td> <td><p>This method now returns a reference to <code>writable</code>.</p></td></tr> <tr><td>v8.0.0</td> <td><p>The <code>chunk</code> argument can now be a <code>Uint8Array</code> instance.</p></td></tr> <tr><td>v0.9.4</td> <td><p><span>Added in: v0.9.4</span></p></td></tr> </tbody></table> </details> </div> <ul> <li><code>chunk</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> | <a href="buffer.html#buffer_class_buffer" class="type"><Buffer></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Uint8Array" class="type"><Uint8Array></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Data_types" class="type"><any></a> Optional data to write. For streams not operating in object mode, <code>chunk</code> must be a string, <code>Buffer</code> or <code>Uint8Array</code>. For object mode streams, <code>chunk</code> may be any JavaScript value other than <code>null</code>.</li> <li><code>encoding</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> The encoding if <code>chunk</code> is a string</li> <li><code>callback</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> Optional callback for when the stream is finished</li> <li>Returns: <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/this" class="type"><this></a></li> </ul> <p>Calling the <code>writable.end()</code> method signals that no more data will be written to the <a href="#stream_class_stream_writable"><code>Writable</code></a>. The optional <code>chunk</code> and <code>encoding</code> arguments allow one final additional chunk of data to be written immediately before closing the stream. If provided, the optional <code>callback</code> function is attached as a listener for the <a href="#stream_event_finish"><code>'finish'</code></a> event.</p> <p>Calling the <a href="#stream_writable_write_chunk_encoding_callback"><code>stream.write()</code></a> method after calling <a href="#stream_writable_end_chunk_encoding_callback"><code>stream.end()</code></a> will raise an error.</p> <pre><code class="language-js">// write 'hello, ' and then end with 'world!' const fs = require('fs'); const file = fs.createWriteStream('example.txt'); file.write('hello, '); file.end('world!'); // writing more now is not allowed! </code></pre> <h5>writable.setDefaultEncoding(encoding)<span><a class="mark" href="#stream_writable_setdefaultencoding_encoding" id="stream_writable_setdefaultencoding_encoding">#</a></span></h5> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v6.1.0</td> <td><p>This method now returns a reference to <code>writable</code>.</p></td></tr> <tr><td>v0.11.15</td> <td><p><span>Added in: v0.11.15</span></p></td></tr> </tbody></table> </details> </div> <ul> <li><code>encoding</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> The new default encoding</li> <li>Returns: <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/this" class="type"><this></a></li> </ul> <p>The <code>writable.setDefaultEncoding()</code> method sets the default <code>encoding</code> for a <a href="#stream_class_stream_writable"><code>Writable</code></a> stream.</p> <h5>writable.uncork()<span><a class="mark" href="#stream_writable_uncork" id="stream_writable_uncork">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.11.2</span> </div> <p>The <code>writable.uncork()</code> method flushes all data buffered since <a href="#stream_writable_cork"><code>stream.cork()</code></a> was called.</p> <p>When using <a href="#stream_writable_cork"><code>writable.cork()</code></a> and <code>writable.uncork()</code> to manage the buffering of writes to a stream, it is recommended that calls to <code>writable.uncork()</code> be deferred using <code>process.nextTick()</code>. Doing so allows batching of all <code>writable.write()</code> calls that occur within a given Node.js event loop phase.</p> <pre><code class="language-js">stream.cork(); stream.write('some '); stream.write('data '); process.nextTick(() => stream.uncork()); </code></pre> <p>If the <a href="#stream_writable_cork"><code>writable.cork()</code></a> method is called multiple times on a stream, the same number of calls to <code>writable.uncork()</code> must be called to flush the buffered data.</p> <pre><code class="language-js">stream.cork(); stream.write('some '); stream.cork(); stream.write('data '); process.nextTick(() => { stream.uncork(); // The data will not be flushed until uncork() is called a second time. stream.uncork(); }); </code></pre> <p>See also: <a href="#stream_writable_cork"><code>writable.cork()</code></a>.</p> <h5>writable.writable<span><a class="mark" href="#stream_writable_writable" id="stream_writable_writable">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.8.0</span> </div> <ul> <li><a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a></li> </ul> <p>Is <code>true</code> if it is safe to call [<code>writable.write()</code>][].</p> <h5>writable.writableHighWaterMark<span><a class="mark" href="#stream_writable_writablehighwatermark" id="stream_writable_writablehighwatermark">#</a></span></h5> <div class="api_metadata"> <span>Added in: v9.3.0</span> </div> <ul> <li><a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Number_type" class="type"><number></a></li> </ul> <p>Return the value of <code>highWaterMark</code> passed when constructing this <code>Writable</code>.</p> <h5>writable.writableLength<span><a class="mark" href="#stream_writable_writablelength" id="stream_writable_writablelength">#</a></span></h5> <div class="api_metadata"> <span>Added in: v9.4.0</span> </div> <p>This property contains the number of bytes (or objects) in the queue ready to be written. The value provides introspection data regarding the status of the <code>highWaterMark</code>.</p> <h5>writable.write(chunk[, encoding][, callback])<span><a class="mark" href="#stream_writable_write_chunk_encoding_callback" id="stream_writable_write_chunk_encoding_callback">#</a></span></h5> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v8.0.0</td> <td><p>The <code>chunk</code> argument can now be a <code>Uint8Array</code> instance.</p></td></tr> <tr><td>v6.0.0</td> <td><p>Passing <code>null</code> as the <code>chunk</code> parameter will always be considered invalid now, even in object mode.</p></td></tr> <tr><td>v0.9.4</td> <td><p><span>Added in: v0.9.4</span></p></td></tr> </tbody></table> </details> </div> <ul> <li><code>chunk</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> | <a href="buffer.html#buffer_class_buffer" class="type"><Buffer></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Uint8Array" class="type"><Uint8Array></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Data_types" class="type"><any></a> Optional data to write. For streams not operating in object mode, <code>chunk</code> must be a string, <code>Buffer</code> or <code>Uint8Array</code>. For object mode streams, <code>chunk</code> may be any JavaScript value other than <code>null</code>.</li> <li><code>encoding</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> The encoding, if <code>chunk</code> is a string</li> <li><code>callback</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> Callback for when this chunk of data is flushed</li> <li>Returns: <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a> <code>false</code> if the stream wishes for the calling code to wait for the <code>'drain'</code> event to be emitted before continuing to write additional data; otherwise <code>true</code>.</li> </ul> <p>The <code>writable.write()</code> method writes some data to the stream, and calls the supplied <code>callback</code> once the data has been fully handled. If an error occurs, the <code>callback</code> <em>may or may not</em> be called with the error as its first argument. To reliably detect write errors, add a listener for the <code>'error'</code> event.</p> <p>The return value is <code>true</code> if the internal buffer is less than the <code>highWaterMark</code> configured when the stream was created after admitting <code>chunk</code>. If <code>false</code> is returned, further attempts to write data to the stream should stop until the <a href="#stream_event_drain"><code>'drain'</code></a> event is emitted.</p> <p>While a stream is not draining, calls to <code>write()</code> will buffer <code>chunk</code>, and return false. Once all currently buffered chunks are drained (accepted for delivery by the operating system), the <code>'drain'</code> event will be emitted. It is recommended that once <code>write()</code> returns false, no more chunks be written until the <code>'drain'</code> event is emitted. While calling <code>write()</code> on a stream that is not draining is allowed, Node.js will buffer all written chunks until maximum memory usage occurs, at which point it will abort unconditionally. Even before it aborts, high memory usage will cause poor garbage collector performance and high RSS (which is not typically released back to the system, even after the memory is no longer required). Since TCP sockets may never drain if the remote peer does not read the data, writing a socket that is not draining may lead to a remotely exploitable vulnerability.</p> <p>Writing data while the stream is not draining is particularly problematic for a <a href="#stream_class_stream_transform"><code>Transform</code></a>, because the <code>Transform</code> streams are paused by default until they are piped or a <code>'data'</code> or <code>'readable'</code> event handler is added.</p> <p>If the data to be written can be generated or fetched on demand, it is recommended to encapsulate the logic into a <a href="#stream_class_stream_readable"><code>Readable</code></a> and use <a href="#stream_readable_pipe_destination_options"><code>stream.pipe()</code></a>. However, if calling <code>write()</code> is preferred, it is possible to respect backpressure and avoid memory issues using the <a href="#stream_event_drain"><code>'drain'</code></a> event:</p> <pre><code class="language-js">function write(data, cb) { if (!stream.write(data)) { stream.once('drain', cb); } else { process.nextTick(cb); } } // Wait for cb to be called before doing any other write. write('hello', () => { console.log('Write completed, do more writes now.'); }); </code></pre> <p>A <code>Writable</code> stream in object mode will always ignore the <code>encoding</code> argument.</p> <h3>Readable Streams<span><a class="mark" href="#stream_readable_streams" id="stream_readable_streams">#</a></span></h3> <p>Readable streams are an abstraction for a <em>source</em> from which data is consumed.</p> <p>Examples of <code>Readable</code> streams include:</p> <ul> <li><a href="http.html#http_class_http_incomingmessage">HTTP responses, on the client</a></li> <li><a href="http.html#http_class_http_incomingmessage">HTTP requests, on the server</a></li> <li><a href="fs.html#fs_class_fs_readstream">fs read streams</a></li> <li><a href="zlib.html">zlib streams</a></li> <li><a href="crypto.html">crypto streams</a></li> <li><a href="net.html#net_class_net_socket">TCP sockets</a></li> <li><a href="child_process.html#child_process_subprocess_stdout">child process stdout and stderr</a></li> <li><a href="process.html#process_process_stdin"><code>process.stdin</code></a></li> </ul> <p>All <a href="#stream_class_stream_readable"><code>Readable</code></a> streams implement the interface defined by the <code>stream.Readable</code> class.</p> <h4>Two Reading Modes<span><a class="mark" href="#stream_two_reading_modes" id="stream_two_reading_modes">#</a></span></h4> <p><code>Readable</code> streams effectively operate in one of two modes: flowing and paused. These modes are separate from <a href="#stream_object_mode">object mode</a>. A <a href="#stream_class_stream_readable"><code>Readable</code></a> stream can be in object mode or not, regardless of whether it is in flowing mode or paused mode.</p> <ul> <li> <p>In flowing mode, data is read from the underlying system automatically and provided to an application as quickly as possible using events via the <a href="events.html#events_class_eventemitter"><code>EventEmitter</code></a> interface.</p> </li> <li> <p>In paused mode, the <a href="#stream_readable_read_size"><code>stream.read()</code></a> method must be called explicitly to read chunks of data from the stream.</p> </li> </ul> <p>All <a href="#stream_class_stream_readable"><code>Readable</code></a> streams begin in paused mode but can be switched to flowing mode in one of the following ways:</p> <ul> <li>Adding a <a href="#stream_event_data"><code>'data'</code></a> event handler.</li> <li>Calling the <a href="#stream_readable_resume"><code>stream.resume()</code></a> method.</li> <li>Calling the <a href="#stream_readable_pipe_destination_options"><code>stream.pipe()</code></a> method to send the data to a <a href="#stream_class_stream_writable"><code>Writable</code></a>.</li> </ul> <p>The <code>Readable</code> can switch back to paused mode using one of the following:</p> <ul> <li>If there are no pipe destinations, by calling the <a href="#stream_readable_pause"><code>stream.pause()</code></a> method.</li> <li>If there are pipe destinations, by removing all pipe destinations. Multiple pipe destinations may be removed by calling the <a href="#stream_readable_unpipe_destination"><code>stream.unpipe()</code></a> method.</li> </ul> <p>The important concept to remember is that a <code>Readable</code> will not generate data until a mechanism for either consuming or ignoring that data is provided. If the consuming mechanism is disabled or taken away, the <code>Readable</code> will <em>attempt</em> to stop generating the data.</p> <p>For backward compatibility reasons, removing <a href="#stream_event_data"><code>'data'</code></a> event handlers will <strong>not</strong> automatically pause the stream. Also, if there are piped destinations, then calling <a href="#stream_readable_pause"><code>stream.pause()</code></a> will not guarantee that the stream will <em>remain</em> paused once those destinations drain and ask for more data.</p> <p>If a <a href="#stream_class_stream_readable"><code>Readable</code></a> is switched into flowing mode and there are no consumers available to handle the data, that data will be lost. This can occur, for instance, when the <code>readable.resume()</code> method is called without a listener attached to the <code>'data'</code> event, or when a <code>'data'</code> event handler is removed from the stream.</p> <p>Adding a <a href="#stream_event_readable"><code>'readable'</code></a> event handler automatically make the stream to stop flowing, and the data to be consumed via <a href="#stream_readable_read_size"><code>readable.read()</code></a>. If the <a href="#stream_event_readable"><code>'readable'</code></a> event handler is removed, then the stream will start flowing again if there is a <a href="#stream_event_data"><code>'data'</code></a> event handler.</p> <h4>Three States<span><a class="mark" href="#stream_three_states" id="stream_three_states">#</a></span></h4> <p>The "two modes" of operation for a <code>Readable</code> stream are a simplified abstraction for the more complicated internal state management that is happening within the <code>Readable</code> stream implementation.</p> <p>Specifically, at any given point in time, every <code>Readable</code> is in one of three possible states:</p> <ul> <li><code>readable.readableFlowing === null</code></li> <li><code>readable.readableFlowing === false</code></li> <li><code>readable.readableFlowing === true</code></li> </ul> <p>When <code>readable.readableFlowing</code> is <code>null</code>, no mechanism for consuming the stream's data is provided. Therefore, the stream will not generate data. While in this state, attaching a listener for the <code>'data'</code> event, calling the <code>readable.pipe()</code> method, or calling the <code>readable.resume()</code> method will switch <code>readable.readableFlowing</code> to <code>true</code>, causing the <code>Readable</code> to begin actively emitting events as data is generated.</p> <p>Calling <code>readable.pause()</code>, <code>readable.unpipe()</code>, or receiving backpressure will cause the <code>readable.readableFlowing</code> to be set as <code>false</code>, temporarily halting the flowing of events but <em>not</em> halting the generation of data. While in this state, attaching a listener for the <code>'data'</code> event will not switch <code>readable.readableFlowing</code> to <code>true</code>.</p> <pre><code class="language-js">const { PassThrough, Writable } = require('stream'); const pass = new PassThrough(); const writable = new Writable(); pass.pipe(writable); pass.unpipe(writable); // readableFlowing is now false pass.on('data', (chunk) => { console.log(chunk.toString()); }); pass.write('ok'); // will not emit 'data' pass.resume(); // must be called to make stream emit 'data' </code></pre> <p>While <code>readable.readableFlowing</code> is <code>false</code>, data may be accumulating within the stream's internal buffer.</p> <h4>Choose One API Style<span><a class="mark" href="#stream_choose_one_api_style" id="stream_choose_one_api_style">#</a></span></h4> <p>The <code>Readable</code> stream API evolved across multiple Node.js versions and provides multiple methods of consuming stream data. In general, developers should choose <em>one</em> of the methods of consuming data and <em>should never</em> use multiple methods to consume data from a single stream. Specifically, using a combination of <code>on('data')</code>, <code>on('readable')</code>, <code>pipe()</code>, or async iterators could lead to unintuitive behavior.</p> <p>Use of the <code>readable.pipe()</code> method is recommended for most users as it has been implemented to provide the easiest way of consuming stream data. Developers that require more fine-grained control over the transfer and generation of data can use the <a href="events.html#events_class_eventemitter"><code>EventEmitter</code></a> and <code>readable.on('readable')</code>/<code>readable.read()</code> or the <code>readable.pause()</code>/<code>readable.resume()</code> APIs.</p> <h4>Class: stream.Readable<span><a class="mark" href="#stream_class_stream_readable" id="stream_class_stream_readable">#</a></span></h4> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <h5>Event: 'close'<span><a class="mark" href="#stream_event_close_1" id="stream_event_close_1">#</a></span></h5> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v10.0.0</td> <td><p>Add <code>emitClose</code> option to specify if <code>'close'</code> is emitted on destroy.</p></td></tr> <tr><td>v0.9.4</td> <td><p><span>Added in: v0.9.4</span></p></td></tr> </tbody></table> </details> </div> <p>The <code>'close'</code> event is emitted when the stream and any of its underlying resources (a file descriptor, for example) have been closed. The event indicates that no more events will be emitted, and no further computation will occur.</p> <p>A <a href="#stream_class_stream_readable"><code>Readable</code></a> stream will always emit the <code>'close'</code> event if it is created with the <code>emitClose</code> option.</p> <h5>Event: 'data'<span><a class="mark" href="#stream_event_data" id="stream_event_data">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <ul> <li><code>chunk</code> <a href="buffer.html#buffer_class_buffer" class="type"><Buffer></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Data_types" class="type"><any></a> The chunk of data. For streams that are not operating in object mode, the chunk will be either a string or <code>Buffer</code>. For streams that are in object mode, the chunk can be any JavaScript value other than <code>null</code>.</li> </ul> <p>The <code>'data'</code> event is emitted whenever the stream is relinquishing ownership of a chunk of data to a consumer. This may occur whenever the stream is switched in flowing mode by calling <code>readable.pipe()</code>, <code>readable.resume()</code>, or by attaching a listener callback to the <code>'data'</code> event. The <code>'data'</code> event will also be emitted whenever the <code>readable.read()</code> method is called and a chunk of data is available to be returned.</p> <p>Attaching a <code>'data'</code> event listener to a stream that has not been explicitly paused will switch the stream into flowing mode. Data will then be passed as soon as it is available.</p> <p>The listener callback will be passed the chunk of data as a string if a default encoding has been specified for the stream using the <code>readable.setEncoding()</code> method; otherwise the data will be passed as a <code>Buffer</code>.</p> <pre><code class="language-js">const readable = getReadableStreamSomehow(); readable.on('data', (chunk) => { console.log(`Received ${chunk.length} bytes of data.`); }); </code></pre> <h5>Event: 'end'<span><a class="mark" href="#stream_event_end" id="stream_event_end">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <p>The <code>'end'</code> event is emitted when there is no more data to be consumed from the stream.</p> <p>The <code>'end'</code> event <strong>will not be emitted</strong> unless the data is completely consumed. This can be accomplished by switching the stream into flowing mode, or by calling <a href="#stream_readable_read_size"><code>stream.read()</code></a> repeatedly until all data has been consumed.</p> <pre><code class="language-js">const readable = getReadableStreamSomehow(); readable.on('data', (chunk) => { console.log(`Received ${chunk.length} bytes of data.`); }); readable.on('end', () => { console.log('There will be no more data.'); }); </code></pre> <h5>Event: 'error'<span><a class="mark" href="#stream_event_error_1" id="stream_event_error_1">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <ul> <li><a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Error" class="type"><Error></a></li> </ul> <p>The <code>'error'</code> event may be emitted by a <code>Readable</code> implementation at any time. Typically, this may occur if the underlying stream is unable to generate data due to an underlying internal failure, or when a stream implementation attempts to push an invalid chunk of data.</p> <p>The listener callback will be passed a single <code>Error</code> object.</p> <h5>Event: 'readable'<span><a class="mark" href="#stream_event_readable" id="stream_event_readable">#</a></span></h5> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v10.0.0</td> <td><p>The <code>'readable'</code> is always emitted in the next tick after <code>.push()</code> is called</p></td></tr> <tr><td>v10.0.0</td> <td><p>Using <code>'readable'</code> requires calling <code>.read()</code>.</p></td></tr> <tr><td>v0.9.4</td> <td><p><span>Added in: v0.9.4</span></p></td></tr> </tbody></table> </details> </div> <p>The <code>'readable'</code> event is emitted when there is data available to be read from the stream. In some cases, attaching a listener for the <code>'readable'</code> event will cause some amount of data to be read into an internal buffer.</p> <pre><code class="language-javascript">const readable = getReadableStreamSomehow(); readable.on('readable', function() { // there is some data to read now let data; while (data = this.read()) { console.log(data); } }); </code></pre> <p>The <code>'readable'</code> event will also be emitted once the end of the stream data has been reached but before the <code>'end'</code> event is emitted.</p> <p>Effectively, the <code>'readable'</code> event indicates that the stream has new information: either new data is available or the end of the stream has been reached. In the former case, <a href="#stream_readable_read_size"><code>stream.read()</code></a> will return the available data. In the latter case, <a href="#stream_readable_read_size"><code>stream.read()</code></a> will return <code>null</code>. For instance, in the following example, <code>foo.txt</code> is an empty file:</p> <pre><code class="language-js">const fs = require('fs'); const rr = fs.createReadStream('foo.txt'); rr.on('readable', () => { console.log(`readable: ${rr.read()}`); }); rr.on('end', () => { console.log('end'); }); </code></pre> <p>The output of running this script is:</p> <pre><code class="language-txt">$ node test.js readable: null end </code></pre> <p>In general, the <code>readable.pipe()</code> and <code>'data'</code> event mechanisms are easier to understand than the <code>'readable'</code> event. However, handling <code>'readable'</code> might result in increased throughput.</p> <p>If both <code>'readable'</code> and <a href="#stream_event_data"><code>'data'</code></a> are used at the same time, <code>'readable'</code> takes precedence in controlling the flow, i.e. <code>'data'</code> will be emitted only when <a href="#stream_readable_read_size"><code>stream.read()</code></a> is called. The <code>readableFlowing</code> property would become <code>false</code>. If there are <code>'data'</code> listeners when <code>'readable'</code> is removed, the stream will start flowing, i.e. <code>'data'</code> events will be emitted without calling <code>.resume()</code>.</p> <h5>readable.destroy([error])<span><a class="mark" href="#stream_readable_destroy_error" id="stream_readable_destroy_error">#</a></span></h5> <div class="api_metadata"> <span>Added in: v8.0.0</span> </div> <ul> <li><code>error</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Error" class="type"><Error></a> Error which will be passed as payload in <code>'error'</code> event</li> <li>Returns: <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/this" class="type"><this></a></li> </ul> <p>Destroy the stream, and emit <code>'error'</code> and <code>'close'</code>. After this call, the readable stream will release any internal resources and subsequent calls to <code>push()</code> will be ignored. Implementors should not override this method, but instead implement <a href="#stream_readable_destroy_err_callback"><code>readable._destroy()</code></a>.</p> <h5>readable.isPaused()<span><a class="mark" href="#stream_readable_ispaused" id="stream_readable_ispaused">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.11.14</span> </div> <ul> <li>Returns: <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a></li> </ul> <p>The <code>readable.isPaused()</code> method returns the current operating state of the <code>Readable</code>. This is used primarily by the mechanism that underlies the <code>readable.pipe()</code> method. In most typical cases, there will be no reason to use this method directly.</p> <pre><code class="language-js">const readable = new stream.Readable(); readable.isPaused(); // === false readable.pause(); readable.isPaused(); // === true readable.resume(); readable.isPaused(); // === false </code></pre> <h5>readable.pause()<span><a class="mark" href="#stream_readable_pause" id="stream_readable_pause">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <ul> <li>Returns: <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/this" class="type"><this></a></li> </ul> <p>The <code>readable.pause()</code> method will cause a stream in flowing mode to stop emitting <a href="#stream_event_data"><code>'data'</code></a> events, switching out of flowing mode. Any data that becomes available will remain in the internal buffer.</p> <pre><code class="language-js">const readable = getReadableStreamSomehow(); readable.on('data', (chunk) => { console.log(`Received ${chunk.length} bytes of data.`); readable.pause(); console.log('There will be no additional data for 1 second.'); setTimeout(() => { console.log('Now data will start flowing again.'); readable.resume(); }, 1000); }); </code></pre> <p>The <code>readable.pause()</code> method has no effect if there is a <code>'readable'</code> event listener.</p> <h5>readable.pipe(destination[, options])<span><a class="mark" href="#stream_readable_pipe_destination_options" id="stream_readable_pipe_destination_options">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <ul> <li><code>destination</code> <a href="stream.html#stream_class_stream_writable" class="type"><stream.Writable></a> The destination for writing data</li> <li> <p><code>options</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Object" class="type"><Object></a> Pipe options</p> <ul> <li><code>end</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a> End the writer when the reader ends. <strong>Default:</strong> <code>true</code>.</li> </ul> </li> <li>Returns: <a href="stream.html#stream_class_stream_writable" class="type"><stream.Writable></a> The <em>destination</em>, allowing for a chain of pipes if it is a <a href="#stream_class_stream_duplex"><code>Duplex</code></a> or a <a href="#stream_class_stream_transform"><code>Transform</code></a> stream</li> </ul> <p>The <code>readable.pipe()</code> method attaches a <a href="#stream_class_stream_writable"><code>Writable</code></a> stream to the <code>readable</code>, causing it to switch automatically into flowing mode and push all of its data to the attached <a href="#stream_class_stream_writable"><code>Writable</code></a>. The flow of data will be automatically managed so that the destination <code>Writable</code> stream is not overwhelmed by a faster <code>Readable</code> stream.</p> <p>The following example pipes all of the data from the <code>readable</code> into a file named <code>file.txt</code>:</p> <pre><code class="language-js">const fs = require('fs'); const readable = getReadableStreamSomehow(); const writable = fs.createWriteStream('file.txt'); // All the data from readable goes into 'file.txt' readable.pipe(writable); </code></pre> <p>It is possible to attach multiple <code>Writable</code> streams to a single <code>Readable</code> stream.</p> <p>The <code>readable.pipe()</code> method returns a reference to the <em>destination</em> stream making it possible to set up chains of piped streams:</p> <pre><code class="language-js">const fs = require('fs'); const r = fs.createReadStream('file.txt'); const z = zlib.createGzip(); const w = fs.createWriteStream('file.txt.gz'); r.pipe(z).pipe(w); </code></pre> <p>By default, <a href="#stream_writable_end_chunk_encoding_callback"><code>stream.end()</code></a> is called on the destination <code>Writable</code> stream when the source <code>Readable</code> stream emits <a href="#stream_event_end"><code>'end'</code></a>, so that the destination is no longer writable. To disable this default behavior, the <code>end</code> option can be passed as <code>false</code>, causing the destination stream to remain open:</p> <pre><code class="language-js">reader.pipe(writer, { end: false }); reader.on('end', () => { writer.end('Goodbye\n'); }); </code></pre> <p>One important caveat is that if the <code>Readable</code> stream emits an error during processing, the <code>Writable</code> destination <em>is not closed</em> automatically. If an error occurs, it will be necessary to <em>manually</em> close each stream in order to prevent memory leaks.</p> <p>The <a href="process.html#process_process_stderr"><code>process.stderr</code></a> and <a href="process.html#process_process_stdout"><code>process.stdout</code></a> <code>Writable</code> streams are never closed until the Node.js process exits, regardless of the specified options.</p> <h5>readable.read([size])<span><a class="mark" href="#stream_readable_read_size" id="stream_readable_read_size">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <ul> <li><code>size</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Number_type" class="type"><number></a> Optional argument to specify how much data to read.</li> <li>Returns: <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> | <a href="buffer.html#buffer_class_buffer" class="type"><Buffer></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Null_type" class="type"><null></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Data_types" class="type"><any></a></li> </ul> <p>The <code>readable.read()</code> method pulls some data out of the internal buffer and returns it. If no data available to be read, <code>null</code> is returned. By default, the data will be returned as a <code>Buffer</code> object unless an encoding has been specified using the <code>readable.setEncoding()</code> method or the stream is operating in object mode.</p> <p>The optional <code>size</code> argument specifies a specific number of bytes to read. If <code>size</code> bytes are not available to be read, <code>null</code> will be returned <em>unless</em> the stream has ended, in which case all of the data remaining in the internal buffer will be returned.</p> <p>If the <code>size</code> argument is not specified, all of the data contained in the internal buffer will be returned.</p> <p>The <code>size</code> argument must be less than or equal to 1 GB.</p> <p>The <code>readable.read()</code> method should only be called on <code>Readable</code> streams operating in paused mode. In flowing mode, <code>readable.read()</code> is called automatically until the internal buffer is fully drained.</p> <pre><code class="language-js">const readable = getReadableStreamSomehow(); readable.on('readable', () => { let chunk; while (null !== (chunk = readable.read())) { console.log(`Received ${chunk.length} bytes of data.`); } }); </code></pre> <p>Note that the <code>while</code> loop is necessary when processing data with <code>readable.read()</code>. Only after <code>readable.read()</code> returns <code>null</code>, <a href=""><code>'readable'</code></a> will be emitted.</p> <p>A <code>Readable</code> stream in object mode will always return a single item from a call to <a href="#stream_readable_read_size"><code>readable.read(size)</code></a>, regardless of the value of the <code>size</code> argument.</p> <p>If the <code>readable.read()</code> method returns a chunk of data, a <code>'data'</code> event will also be emitted.</p> <p>Calling <a href="#stream_readable_read_size"><code>stream.read([size])</code></a> after the <a href="#stream_event_end"><code>'end'</code></a> event has been emitted will return <code>null</code>. No runtime error will be raised.</p> <h5>readable.readable<span><a class="mark" href="#stream_readable_readable" id="stream_readable_readable">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.8.0</span> </div> <ul> <li><a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a></li> </ul> <p>Is <code>true</code> if it is safe to call [<code>readable.read()</code>][].</p> <h5>readable.readableFlowing<span><a class="mark" href="#stream_readable_readableflowing" id="stream_readable_readableflowing">#</a></span></h5> <div class="api_metadata"> <span>Added in: v9.4.0</span> </div> <ul> <li><a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a></li> </ul> <p>This property reflects the current state of a <code>Readable</code> stream as described in the <a href="#stream_three_states">Stream Three States</a> section.</p> <h5>readable.readableHighWaterMark<span><a class="mark" href="#stream_readable_readablehighwatermark" id="stream_readable_readablehighwatermark">#</a></span></h5> <div class="api_metadata"> <span>Added in: v9.3.0</span> </div> <ul> <li><a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Number_type" class="type"><number></a></li> </ul> <p>Returns the value of <code>highWaterMark</code> passed when constructing this <code>Readable</code>.</p> <h5>readable.readableLength<span><a class="mark" href="#stream_readable_readablelength" id="stream_readable_readablelength">#</a></span></h5> <div class="api_metadata"> <span>Added in: v9.4.0</span> </div> <ul> <li><a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Number_type" class="type"><number></a></li> </ul> <p>This property contains the number of bytes (or objects) in the queue ready to be read. The value provides introspection data regarding the status of the <code>highWaterMark</code>.</p> <h5>readable.resume()<span><a class="mark" href="#stream_readable_resume" id="stream_readable_resume">#</a></span></h5> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v10.0.0</td> <td><p>The <code>resume()</code> has no effect if there is a <code>'readable'</code> event listening.</p></td></tr> <tr><td>v0.9.4</td> <td><p><span>Added in: v0.9.4</span></p></td></tr> </tbody></table> </details> </div> <ul> <li>Returns: <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/this" class="type"><this></a></li> </ul> <p>The <code>readable.resume()</code> method causes an explicitly paused <code>Readable</code> stream to resume emitting <a href="#stream_event_data"><code>'data'</code></a> events, switching the stream into flowing mode.</p> <p>The <code>readable.resume()</code> method can be used to fully consume the data from a stream without actually processing any of that data:</p> <pre><code class="language-js">getReadableStreamSomehow() .resume() .on('end', () => { console.log('Reached the end, but did not read anything.'); }); </code></pre> <p>The <code>readable.resume()</code> method has no effect if there is a <code>'readable'</code> event listener.</p> <h5>readable.setEncoding(encoding)<span><a class="mark" href="#stream_readable_setencoding_encoding" id="stream_readable_setencoding_encoding">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <ul> <li><code>encoding</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> The encoding to use.</li> <li>Returns: <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/this" class="type"><this></a></li> </ul> <p>The <code>readable.setEncoding()</code> method sets the character encoding for data read from the <code>Readable</code> stream.</p> <p>By default, no encoding is assigned and stream data will be returned as <code>Buffer</code> objects. Setting an encoding causes the stream data to be returned as strings of the specified encoding rather than as <code>Buffer</code> objects. For instance, calling <code>readable.setEncoding('utf8')</code> will cause the output data to be interpreted as UTF-8 data, and passed as strings. Calling <code>readable.setEncoding('hex')</code> will cause the data to be encoded in hexadecimal string format.</p> <p>The <code>Readable</code> stream will properly handle multi-byte characters delivered through the stream that would otherwise become improperly decoded if simply pulled from the stream as <code>Buffer</code> objects.</p> <pre><code class="language-js">const readable = getReadableStreamSomehow(); readable.setEncoding('utf8'); readable.on('data', (chunk) => { assert.equal(typeof chunk, 'string'); console.log('Got %d characters of string data:', chunk.length); }); </code></pre> <h5>readable.unpipe([destination])<span><a class="mark" href="#stream_readable_unpipe_destination" id="stream_readable_unpipe_destination">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <ul> <li><code>destination</code> <a href="stream.html#stream_class_stream_writable" class="type"><stream.Writable></a> Optional specific stream to unpipe</li> <li>Returns: <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/this" class="type"><this></a></li> </ul> <p>The <code>readable.unpipe()</code> method detaches a <code>Writable</code> stream previously attached using the <a href="#stream_readable_pipe_destination_options"><code>stream.pipe()</code></a> method.</p> <p>If the <code>destination</code> is not specified, then <em>all</em> pipes are detached.</p> <p>If the <code>destination</code> is specified, but no pipe is set up for it, then the method does nothing.</p> <pre><code class="language-js">const fs = require('fs'); const readable = getReadableStreamSomehow(); const writable = fs.createWriteStream('file.txt'); // All the data from readable goes into 'file.txt', // but only for the first second readable.pipe(writable); setTimeout(() => { console.log('Stop writing to file.txt.'); readable.unpipe(writable); console.log('Manually close the file stream.'); writable.end(); }, 1000); </code></pre> <h5>readable.unshift(chunk)<span><a class="mark" href="#stream_readable_unshift_chunk" id="stream_readable_unshift_chunk">#</a></span></h5> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v8.0.0</td> <td><p>The <code>chunk</code> argument can now be a <code>Uint8Array</code> instance.</p></td></tr> <tr><td>v0.9.11</td> <td><p><span>Added in: v0.9.11</span></p></td></tr> </tbody></table> </details> </div> <ul> <li><code>chunk</code> <a href="buffer.html#buffer_class_buffer" class="type"><Buffer></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Uint8Array" class="type"><Uint8Array></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Data_types" class="type"><any></a> Chunk of data to unshift onto the read queue. For streams not operating in object mode, <code>chunk</code> must be a string, <code>Buffer</code> or <code>Uint8Array</code>. For object mode streams, <code>chunk</code> may be any JavaScript value other than <code>null</code>.</li> </ul> <p>The <code>readable.unshift()</code> method pushes a chunk of data back into the internal buffer. This is useful in certain situations where a stream is being consumed by code that needs to "un-consume" some amount of data that it has optimistically pulled out of the source, so that the data can be passed on to some other party.</p> <p>The <code>stream.unshift(chunk)</code> method cannot be called after the <a href="#stream_event_end"><code>'end'</code></a> event has been emitted or a runtime error will be thrown.</p> <p>Developers using <code>stream.unshift()</code> often should consider switching to use of a <a href="#stream_class_stream_transform"><code>Transform</code></a> stream instead. See the <a href="#stream_api_for_stream_implementers">API for Stream Implementers</a> section for more information.</p> <pre><code class="language-js">// Pull off a header delimited by \n\n // use unshift() if we get too much // Call the callback with (error, header, stream) const { StringDecoder } = require('string_decoder'); function parseHeader(stream, callback) { stream.on('error', callback); stream.on('readable', onReadable); const decoder = new StringDecoder('utf8'); let header = ''; function onReadable() { let chunk; while (null !== (chunk = stream.read())) { const str = decoder.write(chunk); if (str.match(/\n\n/)) { // found the header boundary const split = str.split(/\n\n/); header += split.shift(); const remaining = split.join('\n\n'); const buf = Buffer.from(remaining, 'utf8'); stream.removeListener('error', callback); // remove the 'readable' listener before unshifting stream.removeListener('readable', onReadable); if (buf.length) stream.unshift(buf); // now the body of the message can be read from the stream. callback(null, header, stream); } else { // still reading the header. header += str; } } } } </code></pre> <p>Unlike <a href="#stream_readable_push_chunk_encoding"><code>stream.push(chunk)</code></a>, <code>stream.unshift(chunk)</code> will not end the reading process by resetting the internal reading state of the stream. This can cause unexpected results if <code>readable.unshift()</code> is called during a read (i.e. from within a <a href="#stream_readable_read_size_1"><code>stream._read()</code></a> implementation on a custom stream). Following the call to <code>readable.unshift()</code> with an immediate <a href="#stream_readable_push_chunk_encoding"><code>stream.push('')</code></a> will reset the reading state appropriately, however it is best to simply avoid calling <code>readable.unshift()</code> while in the process of performing a read.</p> <h5>readable.wrap(stream)<span><a class="mark" href="#stream_readable_wrap_stream" id="stream_readable_wrap_stream">#</a></span></h5> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <ul> <li><code>stream</code> <a href="stream.html#stream_stream" class="type"><Stream></a> An "old style" readable stream</li> <li>Returns: <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/this" class="type"><this></a></li> </ul> <p>Prior to Node.js 0.10, streams did not implement the entire <code>stream</code> module API as it is currently defined. (See <a href="#stream_compatibility_with_older_node_js_versions">Compatibility</a> for more information.)</p> <p>When using an older Node.js library that emits <a href="#stream_event_data"><code>'data'</code></a> events and has a <a href="#stream_readable_pause"><code>stream.pause()</code></a> method that is advisory only, the <code>readable.wrap()</code> method can be used to create a <a href="#stream_class_stream_readable"><code>Readable</code></a> stream that uses the old stream as its data source.</p> <p>It will rarely be necessary to use <code>readable.wrap()</code> but the method has been provided as a convenience for interacting with older Node.js applications and libraries.</p> <pre><code class="language-js">const { OldReader } = require('./old-api-module.js'); const { Readable } = require('stream'); const oreader = new OldReader(); const myReader = new Readable().wrap(oreader); myReader.on('readable', () => { myReader.read(); // etc. }); </code></pre> <h5>readable[Symbol.asyncIterator]()<span><a class="mark" href="#stream_readable_symbol_asynciterator" id="stream_readable_symbol_asynciterator">#</a></span></h5> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v10.17.0</td> <td><p>Symbol.asyncIterator support is no longer experimental.</p></td></tr> <tr><td>v10.0.0</td> <td><p><span>Added in: v10.0.0</span></p></td></tr> </tbody></table> </details> </div> <p></p><div class="api_stability api_stability_2"><a href="documentation.html#documentation_stability_index">Stability: 2</a> - Stable</div><p></p> <ul> <li>Returns: <a href="https://tc39.github.io/ecma262/#sec-asynciterator-interface" class="type"><AsyncIterator></a> to fully consume the stream.</li> </ul> <pre><code class="language-js">const fs = require('fs'); async function print(readable) { readable.setEncoding('utf8'); let data = ''; for await (const k of readable) { data += k; } console.log(data); } print(fs.createReadStream('file')).catch(console.log); </code></pre> <p>If the loop terminates with a <code>break</code> or a <code>throw</code>, the stream will be destroyed. In other terms, iterating over a stream will consume the stream fully. The stream will be read in chunks of size equal to the <code>highWaterMark</code> option. In the code example above, data will be in a single chunk if the file has less then 64kb of data because no <code>highWaterMark</code> option is provided to <a href="fs.html#fs_fs_createreadstream_path_options"><code>fs.createReadStream()</code></a>.</p> <h3>Duplex and Transform Streams<span><a class="mark" href="#stream_duplex_and_transform_streams" id="stream_duplex_and_transform_streams">#</a></span></h3> <h4>Class: stream.Duplex<span><a class="mark" href="#stream_class_stream_duplex" id="stream_class_stream_duplex">#</a></span></h4> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v6.8.0</td> <td><p>Instances of <code>Duplex</code> now return <code>true</code> when checking <code>instanceof stream.Writable</code>.</p></td></tr> <tr><td>v0.9.4</td> <td><p><span>Added in: v0.9.4</span></p></td></tr> </tbody></table> </details> </div> <p>Duplex streams are streams that implement both the <a href="#stream_class_stream_readable"><code>Readable</code></a> and <a href="#stream_class_stream_writable"><code>Writable</code></a> interfaces.</p> <p>Examples of <code>Duplex</code> streams include:</p> <ul> <li><a href="net.html#net_class_net_socket">TCP sockets</a></li> <li><a href="zlib.html">zlib streams</a></li> <li><a href="crypto.html">crypto streams</a></li> </ul> <h4>Class: stream.Transform<span><a class="mark" href="#stream_class_stream_transform" id="stream_class_stream_transform">#</a></span></h4> <div class="api_metadata"> <span>Added in: v0.9.4</span> </div> <p>Transform streams are <a href="#stream_class_stream_duplex"><code>Duplex</code></a> streams where the output is in some way related to the input. Like all <a href="#stream_class_stream_duplex"><code>Duplex</code></a> streams, <code>Transform</code> streams implement both the <a href="#stream_class_stream_readable"><code>Readable</code></a> and <a href="#stream_class_stream_writable"><code>Writable</code></a> interfaces.</p> <p>Examples of <code>Transform</code> streams include:</p> <ul> <li><a href="zlib.html">zlib streams</a></li> <li><a href="crypto.html">crypto streams</a></li> </ul> <h5>transform.destroy([error])<span><a class="mark" href="#stream_transform_destroy_error" id="stream_transform_destroy_error">#</a></span></h5> <div class="api_metadata"> <span>Added in: v8.0.0</span> </div> <ul> <li><code>error</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Error" class="type"><Error></a></li> </ul> <p>Destroy the stream, and emit <code>'error'</code>. After this call, the transform stream would release any internal resources. Implementors should not override this method, but instead implement <a href="#stream_readable_destroy_err_callback"><code>readable._destroy()</code></a>. The default implementation of <code>_destroy()</code> for <code>Transform</code> also emit <code>'close'</code>.</p> <h3>stream.finished(stream, callback)<span><a class="mark" href="#stream_stream_finished_stream_callback" id="stream_stream_finished_stream_callback">#</a></span></h3> <div class="api_metadata"> <span>Added in: v10.0.0</span> </div> <ul> <li><code>stream</code> <a href="stream.html#stream_stream" class="type"><Stream></a> A readable and/or writable stream.</li> <li><code>callback</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> A callback function that takes an optional error argument.</li> </ul> <p>A function to get notified when a stream is no longer readable, writable or has experienced an error or a premature close event.</p> <pre><code class="language-js">const { finished } = require('stream'); const rs = fs.createReadStream('archive.tar'); finished(rs, (err) => { if (err) { console.error('Stream failed.', err); } else { console.log('Stream is done reading.'); } }); rs.resume(); // drain the stream </code></pre> <p>Especially useful in error handling scenarios where a stream is destroyed prematurely (like an aborted HTTP request), and will not emit <code>'end'</code> or <code>'finish'</code>.</p> <p>The <code>finished</code> API is promisify-able as well;</p> <pre><code class="language-js">const finished = util.promisify(stream.finished); const rs = fs.createReadStream('archive.tar'); async function run() { await finished(rs); console.log('Stream is done reading.'); } run().catch(console.error); rs.resume(); // drain the stream </code></pre> <h3>stream.pipeline(...streams[, callback])<span><a class="mark" href="#stream_stream_pipeline_streams_callback" id="stream_stream_pipeline_streams_callback">#</a></span></h3> <div class="api_metadata"> <span>Added in: v10.0.0</span> </div> <ul> <li><code>...streams</code> <a href="stream.html#stream_stream" class="type"><Stream></a> Two or more streams to pipe between.</li> <li><code>callback</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> A callback function that takes an optional error argument.</li> </ul> <p>A module method to pipe between streams forwarding errors and properly cleaning up and provide a callback when the pipeline is complete.</p> <pre><code class="language-js">const { pipeline } = require('stream'); const fs = require('fs'); const zlib = require('zlib'); // Use the pipeline API to easily pipe a series of streams // together and get notified when the pipeline is fully done. // A pipeline to gzip a potentially huge tar file efficiently: pipeline( fs.createReadStream('archive.tar'), zlib.createGzip(), fs.createWriteStream('archive.tar.gz'), (err) => { if (err) { console.error('Pipeline failed.', err); } else { console.log('Pipeline succeeded.'); } } ); </code></pre> <p>The <code>pipeline</code> API is promisify-able as well:</p> <pre><code class="language-js">const pipeline = util.promisify(stream.pipeline); async function run() { await pipeline( fs.createReadStream('archive.tar'), zlib.createGzip(), fs.createWriteStream('archive.tar.gz') ); console.log('Pipeline succeeded.'); } run().catch(console.error); </code></pre> <h3>Readable.from(iterable, [options])<span><a class="mark" href="#stream_readable_from_iterable_options" id="stream_readable_from_iterable_options">#</a></span></h3> <div class="api_metadata"> <span>Added in: v12.3.0, v10.17.0</span> </div> <ul> <li><code>iterable</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Iteration_protocols#The_iterable_protocol" class="type"><Iterable></a> Object implementing the <code>Symbol.asyncIterator</code> or <code>Symbol.iterator</code> iterable protocol.</li> <li><code>options</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Object" class="type"><Object></a> Options provided to <code>new stream.Readable([options])</code>. By default, <code>Readable.from()</code> will set <code>options.objectMode</code> to <code>true</code>, unless this is explicitly opted out by setting <code>options.objectMode</code> to <code>false</code>.</li> </ul> <p>A utility method for creating Readable Streams out of iterators.</p> <pre><code class="language-js">const { Readable } = require('stream'); async function * generate() { yield 'hello'; yield 'streams'; } const readable = Readable.from(generate()); readable.on('data', (chunk) => { console.log(chunk); }); </code></pre> <h2>API for Stream Implementers<span><a class="mark" href="#stream_api_for_stream_implementers" id="stream_api_for_stream_implementers">#</a></span></h2> <p>The <code>stream</code> module API has been designed to make it possible to easily implement streams using JavaScript's prototypal inheritance model.</p> <p>First, a stream developer would declare a new JavaScript class that extends one of the four basic stream classes (<code>stream.Writable</code>, <code>stream.Readable</code>, <code>stream.Duplex</code>, or <code>stream.Transform</code>), making sure they call the appropriate parent class constructor:</p> <!-- eslint-disable no-useless-constructor --> <pre><code class="language-js">const { Writable } = require('stream'); class MyWritable extends Writable { constructor(options) { super(options); // ... } } </code></pre> <p>The new stream class must then implement one or more specific methods, depending on the type of stream being created, as detailed in the chart below:</p> <table><thead><tr><th>Use-case</th><th>Class</th><th>Method(s) to implement</th></tr></thead><tbody><tr><td>Reading only</td><td><a href="#stream_class_stream_readable"><code>Readable</code></a></td><td><code><a href="#stream_readable_read_size_1">_read</a></code></td></tr><tr><td>Writing only</td><td><a href="#stream_class_stream_writable"><code>Writable</code></a></td><td><code><a href="#stream_writable_write_chunk_encoding_callback_1">_write</a></code>, <code><a href="#stream_writable_writev_chunks_callback">_writev</a></code>, <code><a href="#stream_writable_final_callback">_final</a></code></td></tr><tr><td>Reading and writing</td><td><a href="#stream_class_stream_duplex"><code>Duplex</code></a></td><td><code><a href="#stream_readable_read_size_1">_read</a></code>, <code><a href="#stream_writable_write_chunk_encoding_callback_1">_write</a></code>, <code><a href="#stream_writable_writev_chunks_callback">_writev</a></code>, <code><a href="#stream_writable_final_callback">_final</a></code></td></tr><tr><td>Operate on written data, then read the result</td><td><a href="#stream_class_stream_transform"><code>Transform</code></a></td><td><code><a href="#stream_transform_transform_chunk_encoding_callback">_transform</a></code>, <code><a href="#stream_transform_flush_callback">_flush</a></code>, <code><a href="#stream_writable_final_callback">_final</a></code></td></tr></tbody></table> <p>The implementation code for a stream should <em>never</em> call the "public" methods of a stream that are intended for use by consumers (as described in the <a href="#stream_api_for_stream_consumers">API for Stream Consumers</a> section). Doing so may lead to adverse side effects in application code consuming the stream.</p> <h3>Simplified Construction<span><a class="mark" href="#stream_simplified_construction" id="stream_simplified_construction">#</a></span></h3> <div class="api_metadata"> <span>Added in: v1.2.0</span> </div> <p>For many simple cases, it is possible to construct a stream without relying on inheritance. This can be accomplished by directly creating instances of the <code>stream.Writable</code>, <code>stream.Readable</code>, <code>stream.Duplex</code> or <code>stream.Transform</code> objects and passing appropriate methods as constructor options.</p> <pre><code class="language-js">const { Writable } = require('stream'); const myWritable = new Writable({ write(chunk, encoding, callback) { // ... } }); </code></pre> <h3>Implementing a Writable Stream<span><a class="mark" href="#stream_implementing_a_writable_stream" id="stream_implementing_a_writable_stream">#</a></span></h3> <p>The <code>stream.Writable</code> class is extended to implement a <a href="#stream_class_stream_writable"><code>Writable</code></a> stream.</p> <p>Custom <code>Writable</code> streams <em>must</em> call the <code>new stream.Writable([options])</code> constructor and implement the <code>writable._write()</code> method. The <code>writable._writev()</code> method <em>may</em> also be implemented.</p> <h4>Constructor: new stream.Writable([options])<span><a class="mark" href="#stream_constructor_new_stream_writable_options" id="stream_constructor_new_stream_writable_options">#</a></span></h4> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v10.16.0</td> <td><p>Add <code>autoDestroy</code> option to automatically <code>destroy()</code> the stream when it emits <code>'finish'</code> or errors</p></td></tr> <tr><td>v10.0.0</td> <td><p>Add <code>emitClose</code> option to specify if <code>'close'</code> is emitted on destroy</p></td></tr> </tbody></table> </details> </div> <ul> <li> <p><code>options</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Object" class="type"><Object></a></p> <ul> <li><code>highWaterMark</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Number_type" class="type"><number></a> Buffer level when <a href="#stream_writable_write_chunk_encoding_callback"><code>stream.write()</code></a> starts returning <code>false</code>. <strong>Default:</strong> <code>16384</code> (16kb), or <code>16</code> for <code>objectMode</code> streams.</li> <li><code>decodeStrings</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a> Whether to encode <code>string</code>s passed to <a href="#stream_writable_write_chunk_encoding_callback"><code>stream.write()</code></a> to <code>Buffer</code>s (with the encoding specified in the <a href="#stream_writable_write_chunk_encoding_callback"><code>stream.write()</code></a> call) before passing them to <a href="#stream_writable_write_chunk_encoding_callback_1"><code>stream._write()</code></a>. Other types of data are not converted (i.e. <code>Buffer</code>s are not decoded into <code>string</code>s). Setting to false will prevent <code>string</code>s from being converted. <strong>Default:</strong> <code>true</code>.</li> <li><code>defaultEncoding</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> The default encoding that is used when no encoding is specified as an argument to <a href="#stream_writable_write_chunk_encoding_callback"><code>stream.write()</code></a>. <strong>Default:</strong> <code>'utf8'</code>.</li> <li><code>objectMode</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a> Whether or not the <a href="#stream_writable_write_chunk_encoding_callback"><code>stream.write(anyObj)</code></a> is a valid operation. When set, it becomes possible to write JavaScript values other than string, <code>Buffer</code> or <code>Uint8Array</code> if supported by the stream implementation. <strong>Default:</strong> <code>false</code>.</li> <li><code>emitClose</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a> Whether or not the stream should emit <code>'close'</code> after it has been destroyed. <strong>Default:</strong> <code>true</code>.</li> <li><code>write</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> Implementation for the <a href="#stream_writable_write_chunk_encoding_callback_1"><code>stream._write()</code></a> method.</li> <li><code>writev</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> Implementation for the <a href="#stream_writable_writev_chunks_callback"><code>stream._writev()</code></a> method.</li> <li><code>destroy</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> Implementation for the <a href="#stream_writable_destroy_err_callback"><code>stream._destroy()</code></a> method.</li> <li><code>final</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> Implementation for the <a href="#stream_writable_final_callback"><code>stream._final()</code></a> method.</li> <li><code>autoDestroy</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a> Whether this stream should automatically call <code>.destroy()</code> on itself after ending. <strong>Default:</strong> <code>false</code>.</li> </ul> </li> </ul> <!-- eslint-disable no-useless-constructor --> <pre><code class="language-js">const { Writable } = require('stream'); class MyWritable extends Writable { constructor(options) { // Calls the stream.Writable() constructor super(options); // ... } } </code></pre> <p>Or, when using pre-ES6 style constructors:</p> <pre><code class="language-js">const { Writable } = require('stream'); const util = require('util'); function MyWritable(options) { if (!(this instanceof MyWritable)) return new MyWritable(options); Writable.call(this, options); } util.inherits(MyWritable, Writable); </code></pre> <p>Or, using the Simplified Constructor approach:</p> <pre><code class="language-js">const { Writable } = require('stream'); const myWritable = new Writable({ write(chunk, encoding, callback) { // ... }, writev(chunks, callback) { // ... } }); </code></pre> <h4>writable._write(chunk, encoding, callback)<span><a class="mark" href="#stream_writable_write_chunk_encoding_callback_1" id="stream_writable_write_chunk_encoding_callback_1">#</a></span></h4> <ul> <li><code>chunk</code> <a href="buffer.html#buffer_class_buffer" class="type"><Buffer></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Data_types" class="type"><any></a> The <code>Buffer</code> to be written, converted from the <code>string</code> passed to <a href="#stream_writable_write_chunk_encoding_callback"><code>stream.write()</code></a>. If the stream's <code>decodeStrings</code> option is <code>false</code> or the stream is operating in object mode, the chunk will not be converted & will be whatever was passed to <a href="#stream_writable_write_chunk_encoding_callback"><code>stream.write()</code></a>.</li> <li><code>encoding</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> If the chunk is a string, then <code>encoding</code> is the character encoding of that string. If chunk is a <code>Buffer</code>, or if the stream is operating in object mode, <code>encoding</code> may be ignored.</li> <li><code>callback</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> Call this function (optionally with an error argument) when processing is complete for the supplied chunk.</li> </ul> <p>All <code>Writable</code> stream implementations must provide a <a href="#stream_writable_write_chunk_encoding_callback_1"><code>writable._write()</code></a> method to send data to the underlying resource.</p> <p><a href="#stream_class_stream_transform"><code>Transform</code></a> streams provide their own implementation of the <a href="#stream_writable_write_chunk_encoding_callback_1"><code>writable._write()</code></a>.</p> <p>This function MUST NOT be called by application code directly. It should be implemented by child classes, and called by the internal <code>Writable</code> class methods only.</p> <p>The <code>callback</code> method must be called to signal either that the write completed successfully or failed with an error. The first argument passed to the <code>callback</code> must be the <code>Error</code> object if the call failed or <code>null</code> if the write succeeded.</p> <p>All calls to <code>writable.write()</code> that occur between the time <code>writable._write()</code> is called and the <code>callback</code> is called will cause the written data to be buffered. When the <code>callback</code> is invoked, the stream might emit a <a href="#stream_event_drain"><code>'drain'</code></a> event. If a stream implementation is capable of processing multiple chunks of data at once, the <code>writable._writev()</code> method should be implemented.</p> <p>If the <code>decodeStrings</code> property is explicitly set to <code>false</code> in the constructor options, then <code>chunk</code> will remain the same object that is passed to <code>.write()</code>, and may be a string rather than a <code>Buffer</code>. This is to support implementations that have an optimized handling for certain string data encodings. In that case, the <code>encoding</code> argument will indicate the character encoding of the string. Otherwise, the <code>encoding</code> argument can be safely ignored.</p> <p>The <code>writable._write()</code> method is prefixed with an underscore because it is internal to the class that defines it, and should never be called directly by user programs.</p> <h4>writable._writev(chunks, callback)<span><a class="mark" href="#stream_writable_writev_chunks_callback" id="stream_writable_writev_chunks_callback">#</a></span></h4> <ul> <li><code>chunks</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Object" class="type"><Object[]></a> The chunks to be written. Each chunk has following format: <code>{ chunk: ..., encoding: ... }</code>.</li> <li><code>callback</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> A callback function (optionally with an error argument) to be invoked when processing is complete for the supplied chunks.</li> </ul> <p>This function MUST NOT be called by application code directly. It should be implemented by child classes, and called by the internal <code>Writable</code> class methods only.</p> <p>The <code>writable._writev()</code> method may be implemented in addition to <code>writable._write()</code> in stream implementations that are capable of processing multiple chunks of data at once. If implemented, the method will be called with all chunks of data currently buffered in the write queue.</p> <p>The <code>writable._writev()</code> method is prefixed with an underscore because it is internal to the class that defines it, and should never be called directly by user programs.</p> <h4>writable._destroy(err, callback)<span><a class="mark" href="#stream_writable_destroy_err_callback" id="stream_writable_destroy_err_callback">#</a></span></h4> <div class="api_metadata"> <span>Added in: v8.0.0</span> </div> <ul> <li><code>err</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Error" class="type"><Error></a> A possible error.</li> <li><code>callback</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> A callback function that takes an optional error argument.</li> </ul> <p>The <code>_destroy()</code> method is called by <a href="#stream_writable_destroy_error"><code>writable.destroy()</code></a>. It can be overridden by child classes but it <strong>must not</strong> be called directly.</p> <h4>writable._final(callback)<span><a class="mark" href="#stream_writable_final_callback" id="stream_writable_final_callback">#</a></span></h4> <div class="api_metadata"> <span>Added in: v8.0.0</span> </div> <ul> <li><code>callback</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> Call this function (optionally with an error argument) when finished writing any remaining data.</li> </ul> <p>The <code>_final()</code> method <strong>must not</strong> be called directly. It may be implemented by child classes, and if so, will be called by the internal <code>Writable</code> class methods only.</p> <p>This optional function will be called before the stream closes, delaying the <code>'finish'</code> event until <code>callback</code> is called. This is useful to close resources or write buffered data before a stream ends.</p> <h4>Errors While Writing<span><a class="mark" href="#stream_errors_while_writing" id="stream_errors_while_writing">#</a></span></h4> <p>It is recommended that errors occurring during the processing of the <code>writable._write()</code> and <code>writable._writev()</code> methods are reported by invoking the callback and passing the error as the first argument. This will cause an <code>'error'</code> event to be emitted by the <code>Writable</code>. Throwing an <code>Error</code> from within <code>writable._write()</code> can result in unexpected and inconsistent behavior depending on how the stream is being used. Using the callback ensures consistent and predictable handling of errors.</p> <p>If a <code>Readable</code> stream pipes into a <code>Writable</code> stream when <code>Writable</code> emits an error, the <code>Readable</code> stream will be unpiped.</p> <pre><code class="language-js">const { Writable } = require('stream'); const myWritable = new Writable({ write(chunk, encoding, callback) { if (chunk.toString().indexOf('a') >= 0) { callback(new Error('chunk is invalid')); } else { callback(); } } }); </code></pre> <h4>An Example Writable Stream<span><a class="mark" href="#stream_an_example_writable_stream" id="stream_an_example_writable_stream">#</a></span></h4> <p>The following illustrates a rather simplistic (and somewhat pointless) custom <code>Writable</code> stream implementation. While this specific <code>Writable</code> stream instance is not of any real particular usefulness, the example illustrates each of the required elements of a custom <a href="#stream_class_stream_writable"><code>Writable</code></a> stream instance:</p> <pre><code class="language-js">const { Writable } = require('stream'); class MyWritable extends Writable { _write(chunk, encoding, callback) { if (chunk.toString().indexOf('a') >= 0) { callback(new Error('chunk is invalid')); } else { callback(); } } } </code></pre> <h4>Decoding buffers in a Writable Stream<span><a class="mark" href="#stream_decoding_buffers_in_a_writable_stream" id="stream_decoding_buffers_in_a_writable_stream">#</a></span></h4> <p>Decoding buffers is a common task, for instance, when using transformers whose input is a string. This is not a trivial process when using multi-byte characters encoding, such as UTF-8. The following example shows how to decode multi-byte strings using <code>StringDecoder</code> and <a href="#stream_class_stream_writable"><code>Writable</code></a>.</p> <pre><code class="language-js">const { Writable } = require('stream'); const { StringDecoder } = require('string_decoder'); class StringWritable extends Writable { constructor(options) { super(options); this._decoder = new StringDecoder(options && options.defaultEncoding); this.data = ''; } _write(chunk, encoding, callback) { if (encoding === 'buffer') { chunk = this._decoder.write(chunk); } this.data += chunk; callback(); } _final(callback) { this.data += this._decoder.end(); callback(); } } const euro = [[0xE2, 0x82], [0xAC]].map(Buffer.from); const w = new StringWritable(); w.write('currency: '); w.write(euro[0]); w.end(euro[1]); console.log(w.data); // currency: € </code></pre> <h3>Implementing a Readable Stream<span><a class="mark" href="#stream_implementing_a_readable_stream" id="stream_implementing_a_readable_stream">#</a></span></h3> <p>The <code>stream.Readable</code> class is extended to implement a <a href="#stream_class_stream_readable"><code>Readable</code></a> stream.</p> <p>Custom <code>Readable</code> streams <em>must</em> call the <code>new stream.Readable([options])</code> constructor and implement the <code>readable._read()</code> method.</p> <h4>new stream.Readable([options])<span><a class="mark" href="#stream_new_stream_readable_options" id="stream_new_stream_readable_options">#</a></span></h4> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v10.16.0</td> <td><p>Add <code>autoDestroy</code> option to automatically <code>destroy()</code> the stream when it emits <code>'end'</code> or errors</p></td></tr> </tbody></table> </details> </div> <ul> <li> <p><code>options</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Object" class="type"><Object></a></p> <ul> <li><code>highWaterMark</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Number_type" class="type"><number></a> The maximum <a href="#stream_highwatermark_discrepancy_after_calling_readable_setencoding">number of bytes</a> to store in the internal buffer before ceasing to read from the underlying resource. <strong>Default:</strong> <code>16384</code> (16kb), or <code>16</code> for <code>objectMode</code> streams.</li> <li><code>encoding</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> If specified, then buffers will be decoded to strings using the specified encoding. <strong>Default:</strong> <code>null</code>.</li> <li><code>objectMode</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a> Whether this stream should behave as a stream of objects. Meaning that <a href="#stream_readable_read_size"><code>stream.read(n)</code></a> returns a single value instead of a <code>Buffer</code> of size <code>n</code>. <strong>Default:</strong> <code>false</code>.</li> <li><code>read</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> Implementation for the <a href="#stream_readable_read_size_1"><code>stream._read()</code></a> method.</li> <li><code>destroy</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> Implementation for the <a href="#stream_readable_destroy_err_callback"><code>stream._destroy()</code></a> method.</li> <li><code>autoDestroy</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a> Whether this stream should automatically call <code>.destroy()</code> on itself after ending. <strong>Default:</strong> <code>false</code>.</li> </ul> </li> </ul> <!-- eslint-disable no-useless-constructor --> <pre><code class="language-js">const { Readable } = require('stream'); class MyReadable extends Readable { constructor(options) { // Calls the stream.Readable(options) constructor super(options); // ... } } </code></pre> <p>Or, when using pre-ES6 style constructors:</p> <pre><code class="language-js">const { Readable } = require('stream'); const util = require('util'); function MyReadable(options) { if (!(this instanceof MyReadable)) return new MyReadable(options); Readable.call(this, options); } util.inherits(MyReadable, Readable); </code></pre> <p>Or, using the Simplified Constructor approach:</p> <pre><code class="language-js">const { Readable } = require('stream'); const myReadable = new Readable({ read(size) { // ... } }); </code></pre> <h4>readable._read(size)<span><a class="mark" href="#stream_readable_read_size_1" id="stream_readable_read_size_1">#</a></span></h4> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v10.0.0</td> <td><p>call <code>_read()</code> only once per microtick</p></td></tr> <tr><td>v0.9.4</td> <td><p><span>Added in: v0.9.4</span></p></td></tr> </tbody></table> </details> </div> <ul> <li><code>size</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Number_type" class="type"><number></a> Number of bytes to read asynchronously</li> </ul> <p>This function MUST NOT be called by application code directly. It should be implemented by child classes, and called by the internal <code>Readable</code> class methods only.</p> <p>All <code>Readable</code> stream implementations must provide an implementation of the <code>readable._read()</code> method to fetch data from the underlying resource.</p> <p>When <code>readable._read()</code> is called, if data is available from the resource, the implementation should begin pushing that data into the read queue using the <a href="#stream_readable_push_chunk_encoding"><code>this.push(dataChunk)</code></a> method. <code>_read()</code> should continue reading from the resource and pushing data until <code>readable.push()</code> returns <code>false</code>. Only when <code>_read()</code> is called again after it has stopped should it resume pushing additional data onto the queue.</p> <p>Once the <code>readable._read()</code> method has been called, it will not be called again until the <a href="#stream_readable_push_chunk_encoding"><code>readable.push()</code></a> method is called. <code>readable._read()</code> is guaranteed to be called only once within a synchronous execution, i.e. a microtick.</p> <p>The <code>size</code> argument is advisory. For implementations where a "read" is a single operation that returns data can use the <code>size</code> argument to determine how much data to fetch. Other implementations may ignore this argument and simply provide data whenever it becomes available. There is no need to "wait" until <code>size</code> bytes are available before calling <a href="#stream_readable_push_chunk_encoding"><code>stream.push(chunk)</code></a>.</p> <p>The <code>readable._read()</code> method is prefixed with an underscore because it is internal to the class that defines it, and should never be called directly by user programs.</p> <h4>readable._destroy(err, callback)<span><a class="mark" href="#stream_readable_destroy_err_callback" id="stream_readable_destroy_err_callback">#</a></span></h4> <div class="api_metadata"> <span>Added in: v8.0.0</span> </div> <ul> <li><code>err</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Error" class="type"><Error></a> A possible error.</li> <li><code>callback</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> A callback function that takes an optional error argument.</li> </ul> <p>The <code>_destroy()</code> method is called by <a href="#stream_readable_destroy_error"><code>readable.destroy()</code></a>. It can be overridden by child classes but it <strong>must not</strong> be called directly.</p> <h4>readable.push(chunk[, encoding])<span><a class="mark" href="#stream_readable_push_chunk_encoding" id="stream_readable_push_chunk_encoding">#</a></span></h4> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v8.0.0</td> <td><p>The <code>chunk</code> argument can now be a <code>Uint8Array</code> instance.</p></td></tr> </tbody></table> </details> </div> <ul> <li><code>chunk</code> <a href="buffer.html#buffer_class_buffer" class="type"><Buffer></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Uint8Array" class="type"><Uint8Array></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Null_type" class="type"><null></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Data_types" class="type"><any></a> Chunk of data to push into the read queue. For streams not operating in object mode, <code>chunk</code> must be a string, <code>Buffer</code> or <code>Uint8Array</code>. For object mode streams, <code>chunk</code> may be any JavaScript value.</li> <li><code>encoding</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> Encoding of string chunks. Must be a valid <code>Buffer</code> encoding, such as <code>'utf8'</code> or <code>'ascii'</code>.</li> <li>Returns: <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a> <code>true</code> if additional chunks of data may continue to be pushed; <code>false</code> otherwise.</li> </ul> <p>When <code>chunk</code> is a <code>Buffer</code>, <code>Uint8Array</code> or <code>string</code>, the <code>chunk</code> of data will be added to the internal queue for users of the stream to consume. Passing <code>chunk</code> as <code>null</code> signals the end of the stream (EOF), after which no more data can be written.</p> <p>When the <code>Readable</code> is operating in paused mode, the data added with <code>readable.push()</code> can be read out by calling the <a href="#stream_readable_read_size"><code>readable.read()</code></a> method when the <a href="#stream_event_readable"><code>'readable'</code></a> event is emitted.</p> <p>When the <code>Readable</code> is operating in flowing mode, the data added with <code>readable.push()</code> will be delivered by emitting a <code>'data'</code> event.</p> <p>The <code>readable.push()</code> method is designed to be as flexible as possible. For example, when wrapping a lower-level source that provides some form of pause/resume mechanism, and a data callback, the low-level source can be wrapped by the custom <code>Readable</code> instance:</p> <pre><code class="language-js">// source is an object with readStop() and readStart() methods, // and an `ondata` member that gets called when it has data, and // an `onend` member that gets called when the data is over. class SourceWrapper extends Readable { constructor(options) { super(options); this._source = getLowlevelSourceObject(); // Every time there's data, push it into the internal buffer. this._source.ondata = (chunk) => { // if push() returns false, then stop reading from source if (!this.push(chunk)) this._source.readStop(); }; // When the source ends, push the EOF-signaling `null` chunk this._source.onend = () => { this.push(null); }; } // _read will be called when the stream wants to pull more data in // the advisory size argument is ignored in this case. _read(size) { this._source.readStart(); } } </code></pre> <p>The <code>readable.push()</code> method is intended be called only by <code>Readable</code> implementers, and only from within the <code>readable._read()</code> method.</p> <p>For streams not operating in object mode, if the <code>chunk</code> parameter of <code>readable.push()</code> is <code>undefined</code>, it will be treated as empty string or buffer. See <a href="#stream_readable_push"><code>readable.push('')</code></a> for more information.</p> <h4>Errors While Reading<span><a class="mark" href="#stream_errors_while_reading" id="stream_errors_while_reading">#</a></span></h4> <p>It is recommended that errors occurring during the processing of the <code>readable._read()</code> method are emitted using the <code>'error'</code> event rather than being thrown. Throwing an <code>Error</code> from within <code>readable._read()</code> can result in unexpected and inconsistent behavior depending on whether the stream is operating in flowing or paused mode. Using the <code>'error'</code> event ensures consistent and predictable handling of errors.</p> <!-- eslint-disable no-useless-return --> <pre><code class="language-js">const { Readable } = require('stream'); const myReadable = new Readable({ read(size) { if (checkSomeErrorCondition()) { process.nextTick(() => this.emit('error', err)); return; } // do some work } }); </code></pre> <h4>An Example Counting Stream<span><a class="mark" href="#stream_an_example_counting_stream" id="stream_an_example_counting_stream">#</a></span></h4> <p>The following is a basic example of a <code>Readable</code> stream that emits the numerals from 1 to 1,000,000 in ascending order, and then ends.</p> <pre><code class="language-js">const { Readable } = require('stream'); class Counter extends Readable { constructor(opt) { super(opt); this._max = 1000000; this._index = 1; } _read() { const i = this._index++; if (i > this._max) this.push(null); else { const str = String(i); const buf = Buffer.from(str, 'ascii'); this.push(buf); } } } </code></pre> <h3>Implementing a Duplex Stream<span><a class="mark" href="#stream_implementing_a_duplex_stream" id="stream_implementing_a_duplex_stream">#</a></span></h3> <p>A <a href="#stream_class_stream_duplex"><code>Duplex</code></a> stream is one that implements both <a href="#stream_class_stream_readable"><code>Readable</code></a> and <a href="#stream_class_stream_writable"><code>Writable</code></a>, such as a TCP socket connection.</p> <p>Because JavaScript does not have support for multiple inheritance, the <code>stream.Duplex</code> class is extended to implement a <a href="#stream_class_stream_duplex"><code>Duplex</code></a> stream (as opposed to extending the <code>stream.Readable</code> <em>and</em> <code>stream.Writable</code> classes).</p> <p>The <code>stream.Duplex</code> class prototypically inherits from <code>stream.Readable</code> and parasitically from <code>stream.Writable</code>, but <code>instanceof</code> will work properly for both base classes due to overriding <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Symbol/hasInstance"><code>Symbol.hasInstance</code></a> on <code>stream.Writable</code>.</p> <p>Custom <code>Duplex</code> streams <em>must</em> call the <code>new stream.Duplex([options])</code> constructor and implement <em>both</em> the <code>readable._read()</code> and <code>writable._write()</code> methods.</p> <h4>new stream.Duplex(options)<span><a class="mark" href="#stream_new_stream_duplex_options" id="stream_new_stream_duplex_options">#</a></span></h4> <div class="api_metadata"> <details class="changelog"><summary>History</summary> <table> <tbody><tr><th>Version</th><th>Changes</th></tr> <tr><td>v8.4.0</td> <td><p>The <code>readableHighWaterMark</code> and <code>writableHighWaterMark</code> options are supported now.</p></td></tr> </tbody></table> </details> </div> <ul> <li> <p><code>options</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Object" class="type"><Object></a> Passed to both <code>Writable</code> and <code>Readable</code> constructors. Also has the following fields:</p> <ul> <li><code>allowHalfOpen</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a> If set to <code>false</code>, then the stream will automatically end the writable side when the readable side ends. <strong>Default:</strong> <code>true</code>.</li> <li><code>readableObjectMode</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a> Sets <code>objectMode</code> for readable side of the stream. Has no effect if <code>objectMode</code> is <code>true</code>. <strong>Default:</strong> <code>false</code>.</li> <li><code>writableObjectMode</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Boolean_type" class="type"><boolean></a> Sets <code>objectMode</code> for writable side of the stream. Has no effect if <code>objectMode</code> is <code>true</code>. <strong>Default:</strong> <code>false</code>.</li> <li><code>readableHighWaterMark</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Number_type" class="type"><number></a> Sets <code>highWaterMark</code> for the readable side of the stream. Has no effect if <code>highWaterMark</code> is provided.</li> <li><code>writableHighWaterMark</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Number_type" class="type"><number></a> Sets <code>highWaterMark</code> for the writable side of the stream. Has no effect if <code>highWaterMark</code> is provided.</li> </ul> </li> </ul> <!-- eslint-disable no-useless-constructor --> <pre><code class="language-js">const { Duplex } = require('stream'); class MyDuplex extends Duplex { constructor(options) { super(options); // ... } } </code></pre> <p>Or, when using pre-ES6 style constructors:</p> <pre><code class="language-js">const { Duplex } = require('stream'); const util = require('util'); function MyDuplex(options) { if (!(this instanceof MyDuplex)) return new MyDuplex(options); Duplex.call(this, options); } util.inherits(MyDuplex, Duplex); </code></pre> <p>Or, using the Simplified Constructor approach:</p> <pre><code class="language-js">const { Duplex } = require('stream'); const myDuplex = new Duplex({ read(size) { // ... }, write(chunk, encoding, callback) { // ... } }); </code></pre> <h4>An Example Duplex Stream<span><a class="mark" href="#stream_an_example_duplex_stream" id="stream_an_example_duplex_stream">#</a></span></h4> <p>The following illustrates a simple example of a <code>Duplex</code> stream that wraps a hypothetical lower-level source object to which data can be written, and from which data can be read, albeit using an API that is not compatible with Node.js streams. The following illustrates a simple example of a <code>Duplex</code> stream that buffers incoming written data via the <a href="#stream_class_stream_writable"><code>Writable</code></a> interface that is read back out via the <a href="#stream_class_stream_readable"><code>Readable</code></a> interface.</p> <pre><code class="language-js">const { Duplex } = require('stream'); const kSource = Symbol('source'); class MyDuplex extends Duplex { constructor(source, options) { super(options); this[kSource] = source; } _write(chunk, encoding, callback) { // The underlying source only deals with strings if (Buffer.isBuffer(chunk)) chunk = chunk.toString(); this[kSource].writeSomeData(chunk); callback(); } _read(size) { this[kSource].fetchSomeData(size, (data, encoding) => { this.push(Buffer.from(data, encoding)); }); } } </code></pre> <p>The most important aspect of a <code>Duplex</code> stream is that the <code>Readable</code> and <code>Writable</code> sides operate independently of one another despite co-existing within a single object instance.</p> <h4>Object Mode Duplex Streams<span><a class="mark" href="#stream_object_mode_duplex_streams" id="stream_object_mode_duplex_streams">#</a></span></h4> <p>For <code>Duplex</code> streams, <code>objectMode</code> can be set exclusively for either the <code>Readable</code> or <code>Writable</code> side using the <code>readableObjectMode</code> and <code>writableObjectMode</code> options respectively.</p> <p>In the following example, for instance, a new <code>Transform</code> stream (which is a type of <a href="#stream_class_stream_duplex"><code>Duplex</code></a> stream) is created that has an object mode <code>Writable</code> side that accepts JavaScript numbers that are converted to hexadecimal strings on the <code>Readable</code> side.</p> <pre><code class="language-js">const { Transform } = require('stream'); // All Transform streams are also Duplex Streams const myTransform = new Transform({ writableObjectMode: true, transform(chunk, encoding, callback) { // Coerce the chunk to a number if necessary chunk |= 0; // Transform the chunk into something else. const data = chunk.toString(16); // Push the data onto the readable queue. callback(null, '0'.repeat(data.length % 2) + data); } }); myTransform.setEncoding('ascii'); myTransform.on('data', (chunk) => console.log(chunk)); myTransform.write(1); // Prints: 01 myTransform.write(10); // Prints: 0a myTransform.write(100); // Prints: 64 </code></pre> <h3>Implementing a Transform Stream<span><a class="mark" href="#stream_implementing_a_transform_stream" id="stream_implementing_a_transform_stream">#</a></span></h3> <p>A <a href="#stream_class_stream_transform"><code>Transform</code></a> stream is a <a href="#stream_class_stream_duplex"><code>Duplex</code></a> stream where the output is computed in some way from the input. Examples include <a href="zlib.html">zlib</a> streams or <a href="crypto.html">crypto</a> streams that compress, encrypt, or decrypt data.</p> <p>There is no requirement that the output be the same size as the input, the same number of chunks, or arrive at the same time. For example, a <code>Hash</code> stream will only ever have a single chunk of output which is provided when the input is ended. A <code>zlib</code> stream will produce output that is either much smaller or much larger than its input.</p> <p>The <code>stream.Transform</code> class is extended to implement a <a href="#stream_class_stream_transform"><code>Transform</code></a> stream.</p> <p>The <code>stream.Transform</code> class prototypically inherits from <code>stream.Duplex</code> and implements its own versions of the <code>writable._write()</code> and <code>readable._read()</code> methods. Custom <code>Transform</code> implementations <em>must</em> implement the <a href="#stream_transform_transform_chunk_encoding_callback"><code>transform._transform()</code></a> method and <em>may</em> also implement the <a href="#stream_transform_flush_callback"><code>transform._flush()</code></a> method.</p> <p>Care must be taken when using <code>Transform</code> streams in that data written to the stream can cause the <code>Writable</code> side of the stream to become paused if the output on the <code>Readable</code> side is not consumed.</p> <h4>new stream.Transform([options])<span><a class="mark" href="#stream_new_stream_transform_options" id="stream_new_stream_transform_options">#</a></span></h4> <ul> <li> <p><code>options</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Object" class="type"><Object></a> Passed to both <code>Writable</code> and <code>Readable</code> constructors. Also has the following fields:</p> <ul> <li><code>transform</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> Implementation for the <a href="#stream_transform_transform_chunk_encoding_callback"><code>stream._transform()</code></a> method.</li> <li><code>flush</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> Implementation for the <a href="#stream_transform_flush_callback"><code>stream._flush()</code></a> method.</li> </ul> </li> </ul> <!-- eslint-disable no-useless-constructor --> <pre><code class="language-js">const { Transform } = require('stream'); class MyTransform extends Transform { constructor(options) { super(options); // ... } } </code></pre> <p>Or, when using pre-ES6 style constructors:</p> <pre><code class="language-js">const { Transform } = require('stream'); const util = require('util'); function MyTransform(options) { if (!(this instanceof MyTransform)) return new MyTransform(options); Transform.call(this, options); } util.inherits(MyTransform, Transform); </code></pre> <p>Or, using the Simplified Constructor approach:</p> <pre><code class="language-js">const { Transform } = require('stream'); const myTransform = new Transform({ transform(chunk, encoding, callback) { // ... } }); </code></pre> <h4>Events: 'finish' and 'end'<span><a class="mark" href="#stream_events_finish_and_end" id="stream_events_finish_and_end">#</a></span></h4> <p>The <a href="#stream_event_finish"><code>'finish'</code></a> and <a href="#stream_event_end"><code>'end'</code></a> events are from the <code>stream.Writable</code> and <code>stream.Readable</code> classes, respectively. The <code>'finish'</code> event is emitted after <a href="#stream_writable_end_chunk_encoding_callback"><code>stream.end()</code></a> is called and all chunks have been processed by <a href="#stream_transform_transform_chunk_encoding_callback"><code>stream._transform()</code></a>. The <code>'end'</code> event is emitted after all data has been output, which occurs after the callback in <a href="#stream_transform_flush_callback"><code>transform._flush()</code></a> has been called.</p> <h4>transform._flush(callback)<span><a class="mark" href="#stream_transform_flush_callback" id="stream_transform_flush_callback">#</a></span></h4> <ul> <li><code>callback</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> A callback function (optionally with an error argument and data) to be called when remaining data has been flushed.</li> </ul> <p>This function MUST NOT be called by application code directly. It should be implemented by child classes, and called by the internal <code>Readable</code> class methods only.</p> <p>In some cases, a transform operation may need to emit an additional bit of data at the end of the stream. For example, a <code>zlib</code> compression stream will store an amount of internal state used to optimally compress the output. When the stream ends, however, that additional data needs to be flushed so that the compressed data will be complete.</p> <p>Custom <a href="#stream_class_stream_transform"><code>Transform</code></a> implementations <em>may</em> implement the <code>transform._flush()</code> method. This will be called when there is no more written data to be consumed, but before the <a href="#stream_event_end"><code>'end'</code></a> event is emitted signaling the end of the <a href="#stream_class_stream_readable"><code>Readable</code></a> stream.</p> <p>Within the <code>transform._flush()</code> implementation, the <code>readable.push()</code> method may be called zero or more times, as appropriate. The <code>callback</code> function must be called when the flush operation is complete.</p> <p>The <code>transform._flush()</code> method is prefixed with an underscore because it is internal to the class that defines it, and should never be called directly by user programs.</p> <h4>transform._transform(chunk, encoding, callback)<span><a class="mark" href="#stream_transform_transform_chunk_encoding_callback" id="stream_transform_transform_chunk_encoding_callback">#</a></span></h4> <ul> <li><code>chunk</code> <a href="buffer.html#buffer_class_buffer" class="type"><Buffer></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> | <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#Data_types" class="type"><any></a> The <code>Buffer</code> to be transformed, converted from the <code>string</code> passed to <a href="#stream_writable_write_chunk_encoding_callback"><code>stream.write()</code></a>. If the stream's <code>decodeStrings</code> option is <code>false</code> or the stream is operating in object mode, the chunk will not be converted & will be whatever was passed to <a href="#stream_writable_write_chunk_encoding_callback"><code>stream.write()</code></a>.</li> <li><code>encoding</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Data_structures#String_type" class="type"><string></a> If the chunk is a string, then this is the encoding type. If chunk is a buffer, then this is the special value - 'buffer', ignore it in this case.</li> <li><code>callback</code> <a href="https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Function" class="type"><Function></a> A callback function (optionally with an error argument and data) to be called after the supplied <code>chunk</code> has been processed.</li> </ul> <p>This function MUST NOT be called by application code directly. It should be implemented by child classes, and called by the internal <code>Readable</code> class methods only.</p> <p>All <code>Transform</code> stream implementations must provide a <code>_transform()</code> method to accept input and produce output. The <code>transform._transform()</code> implementation handles the bytes being written, computes an output, then passes that output off to the readable portion using the <code>readable.push()</code> method.</p> <p>The <code>transform.push()</code> method may be called zero or more times to generate output from a single input chunk, depending on how much is to be output as a result of the chunk.</p> <p>It is possible that no output is generated from any given chunk of input data.</p> <p>The <code>callback</code> function must be called only when the current chunk is completely consumed. The first argument passed to the <code>callback</code> must be an <code>Error</code> object if an error occurred while processing the input or <code>null</code> otherwise. If a second argument is passed to the <code>callback</code>, it will be forwarded on to the <code>readable.push()</code> method. In other words, the following are equivalent:</p> <pre><code class="language-js">transform.prototype._transform = function(data, encoding, callback) { this.push(data); callback(); }; transform.prototype._transform = function(data, encoding, callback) { callback(null, data); }; </code></pre> <p>The <code>transform._transform()</code> method is prefixed with an underscore because it is internal to the class that defines it, and should never be called directly by user programs.</p> <p><code>transform._transform()</code> is never called in parallel; streams implement a queue mechanism, and to receive the next chunk, <code>callback</code> must be called, either synchronously or asynchronously.</p> <h4>Class: stream.PassThrough<span><a class="mark" href="#stream_class_stream_passthrough" id="stream_class_stream_passthrough">#</a></span></h4> <p>The <code>stream.PassThrough</code> class is a trivial implementation of a <a href="#stream_class_stream_transform"><code>Transform</code></a> stream that simply passes the input bytes across to the output. Its purpose is primarily for examples and testing, but there are some use cases where <code>stream.PassThrough</code> is useful as a building block for novel sorts of streams.</p> <h2>Additional Notes<span><a class="mark" href="#stream_additional_notes" id="stream_additional_notes">#</a></span></h2> <h3>Streams Compatibility with Async Generators and Async Iterators<span><a class="mark" href="#stream_streams_compatibility_with_async_generators_and_async_iterators" id="stream_streams_compatibility_with_async_generators_and_async_iterators">#</a></span></h3> <p>With the support of async generators and iterators in JavaScript, async generators are effectively a first-class language-level stream construct at this point.</p> <p>Some common interop cases of using Node.js streams with async generators and async iterators are provided below.</p> <h4>Consuming Readable Streams with Async Iterators<span><a class="mark" href="#stream_consuming_readable_streams_with_async_iterators" id="stream_consuming_readable_streams_with_async_iterators">#</a></span></h4> <pre><code class="language-js">(async function() { for await (const chunk of readable) { console.log(chunk); } })(); </code></pre> <h4>Creating Readable Streams with Async Generators<span><a class="mark" href="#stream_creating_readable_streams_with_async_generators" id="stream_creating_readable_streams_with_async_generators">#</a></span></h4> <p>We can construct a Node.js Readable Stream from an asynchronous generator using the <code>Readable.from</code> utility method:</p> <pre><code class="language-js">const { Readable } = require('stream'); async function * generate() { yield 'a'; yield 'b'; yield 'c'; } const readable = Readable.from(generate()); readable.on('data', (chunk) => { console.log(chunk); }); </code></pre> <h4>Piping to Writable Streams from Async Iterators<span><a class="mark" href="#stream_piping_to_writable_streams_from_async_iterators" id="stream_piping_to_writable_streams_from_async_iterators">#</a></span></h4> <p>In the scenario of writing to a writeable stream from an async iterator, it is important to ensure the correct handling of backpressure and errors.</p> <pre><code class="language-js">const { once } = require('events'); const writeable = fs.createWriteStream('./file'); (async function() { for await (const chunk of iterator) { // Handle backpressure on write if (!writeable.write(value)) await once(writeable, 'drain'); } writeable.end(); // Ensure completion without errors await once(writeable, 'finish'); })(); </code></pre> <p>In the above, errors on the write stream would be caught and thrown by the two <code>once</code> listeners, since <code>once</code> will also handle <code>'error'</code> events.</p> <p>Alternatively the readable stream could be wrapped with <code>Readable.from</code> and then piped via <code>.pipe</code>:</p> <pre><code class="language-js">const { once } = require('events'); const writeable = fs.createWriteStream('./file'); (async function() { const readable = Readable.from(iterator); readable.pipe(writeable); // Ensure completion without errors await once(writeable, 'finish'); })(); </code></pre> <h3>Compatibility with Older Node.js Versions<span><a class="mark" href="#stream_compatibility_with_older_node_js_versions" id="stream_compatibility_with_older_node_js_versions">#</a></span></h3> <p>Prior to Node.js 0.10, the <code>Readable</code> stream interface was simpler, but also less powerful and less useful.</p> <ul> <li>Rather than waiting for calls to the <a href="#stream_readable_read_size"><code>stream.read()</code></a> method, <a href="#stream_event_data"><code>'data'</code></a> events would begin emitting immediately. Applications that would need to perform some amount of work to decide how to handle data were required to store read data into buffers so the data would not be lost.</li> <li>The <a href="#stream_readable_pause"><code>stream.pause()</code></a> method was advisory, rather than guaranteed. This meant that it was still necessary to be prepared to receive <a href="#stream_event_data"><code>'data'</code></a> events <em>even when the stream was in a paused state</em>.</li> </ul> <p>In Node.js 0.10, the <a href="#stream_class_stream_readable"><code>Readable</code></a> class was added. For backward compatibility with older Node.js programs, <code>Readable</code> streams switch into "flowing mode" when a <a href="#stream_event_data"><code>'data'</code></a> event handler is added, or when the <a href="#stream_readable_resume"><code>stream.resume()</code></a> method is called. The effect is that, even when not using the new <a href="#stream_readable_read_size"><code>stream.read()</code></a> method and <a href="#stream_event_readable"><code>'readable'</code></a> event, it is no longer necessary to worry about losing <a href="#stream_event_data"><code>'data'</code></a> chunks.</p> <p>While most applications will continue to function normally, this introduces an edge case in the following conditions:</p> <ul> <li>No <a href="#stream_event_data"><code>'data'</code></a> event listener is added.</li> <li>The <a href="#stream_readable_resume"><code>stream.resume()</code></a> method is never called.</li> <li>The stream is not piped to any writable destination.</li> </ul> <p>For example, consider the following code:</p> <pre><code class="language-js">// WARNING! BROKEN! net.createServer((socket) => { // we add an 'end' listener, but never consume the data socket.on('end', () => { // It will never get here. socket.end('The message was received but was not processed.\n'); }); }).listen(1337); </code></pre> <p>Prior to Node.js 0.10, the incoming message data would be simply discarded. However, in Node.js 0.10 and beyond, the socket remains paused forever.</p> <p>The workaround in this situation is to call the <a href="#stream_readable_resume"><code>stream.resume()</code></a> method to begin the flow of data:</p> <pre><code class="language-js">// Workaround net.createServer((socket) => { socket.on('end', () => { socket.end('The message was received but was not processed.\n'); }); // start the flow of data, discarding it. socket.resume(); }).listen(1337); </code></pre> <p>In addition to new <code>Readable</code> streams switching into flowing mode, pre-0.10 style streams can be wrapped in a <code>Readable</code> class using the <a href="#stream_readable_wrap_stream"><code>readable.wrap()</code></a> method.</p> <h3><code>readable.read(0)</code><span><a class="mark" href="#stream_readable_read_0" id="stream_readable_read_0">#</a></span></h3> <p>There are some cases where it is necessary to trigger a refresh of the underlying readable stream mechanisms, without actually consuming any data. In such cases, it is possible to call <code>readable.read(0)</code>, which will always return <code>null</code>.</p> <p>If the internal read buffer is below the <code>highWaterMark</code>, and the stream is not currently reading, then calling <code>stream.read(0)</code> will trigger a low-level <a href="#stream_readable_read_size_1"><code>stream._read()</code></a> call.</p> <p>While most applications will almost never need to do this, there are situations within Node.js where this is done, particularly in the <code>Readable</code> stream class internals.</p> <h3><code>readable.push('')</code><span><a class="mark" href="#stream_readable_push" id="stream_readable_push">#</a></span></h3> <p>Use of <code>readable.push('')</code> is not recommended.</p> <p>Pushing a zero-byte string, <code>Buffer</code> or <code>Uint8Array</code> to a stream that is not in object mode has an interesting side effect. Because it <em>is</em> a call to <a href="#stream_readable_push_chunk_encoding"><code>readable.push()</code></a>, the call will end the reading process. However, because the argument is an empty string, no data is added to the readable buffer so there is nothing for a user to consume.</p> <h3><code>highWaterMark</code> discrepancy after calling <code>readable.setEncoding()</code><span><a class="mark" href="#stream_highwatermark_discrepancy_after_calling_readable_setencoding" id="stream_highwatermark_discrepancy_after_calling_readable_setencoding">#</a></span></h3> <p>The use of <code>readable.setEncoding()</code> will change the behavior of how the <code>highWaterMark</code> operates in non-object mode.</p> <p>Typically, the size of the current buffer is measured against the <code>highWaterMark</code> in <em>bytes</em>. However, after <code>setEncoding()</code> is called, the comparison function will begin to measure the buffer's size in <em>characters</em>.</p> <p>This is not a problem in common cases with <code>latin1</code> or <code>ascii</code>. But it is advised to be mindful about this behavior when working with strings that could contain multi-byte characters.</p> </div> </div> </div> <script src="assets/sh_main.js"></script> <script src="assets/sh_javascript.min.js"></script> <script>highlight(undefined, undefined, 'pre');</script> </body> </html>