Speed up node.js when operating with rows from large file - javascript

I want to read a file with many rows, then write results.
It's fine with small files <50kb.
But I've got 15MB file for programming competition - as a hard input.
Node.js become slow and I can't get the output in time, because I have to send them the output within few minutes.
And it's even not using full CPU/RAM.
Is the problem in my code, or can I do something about it? Thanks!
const fs = require("fs");
const input = "D:\\Downloads\\example.txt";
const output = input + ".final.txt";
var lineReader = require("readline").createInterface({
input: fs.createReadStream(input),
});
let out = "";
let all = [];
const line_counter = (
(i = 0) =>
() =>
++i
)();
lineReader.on("line", function (radek, index = line_counter()) {
all.push(radek);
});
all.forEach((v) => {
out += `${v}\n`;
});
fs.writeFile(output, out, (err) => {
if (err) {
console.error(err);
}
});

It seems like you want a better understanding of how to do the following using a streaming technique:
read an input text file stream line-by-line
perform a transform operation on each line of text
write the result of each transform operation to an output file stream
Node supports the web standard streams API — see the list of global objects in the current LTS version of Node (18): https://nodejs.org/docs/latest-v18.x/api/globals.html.
Below, I'll include a complete, minimal example which demonstrates the criteria above — you can use it as a model for learning and adapt it to meet the needs of your program. Because your goal is learning, I've included verbose comments at every step of the program, including links to documentation.
You'll also probably find it helpful to read about the Streams API on MDN and web.dev.
module.mjs:
import {open} from 'node:fs/promises';
import {Writable} from 'node:stream';
// Break string chunks from a ReadableStream into lines. Adapted from Deno's std library:
// See: https://github.com/denoland/deno_std/blob/0.166.0/streams/delimiter.ts#L11-L68
class TextLineStream extends TransformStream {
#buf = "";
constructor() {
super({
transform: (chunk, controller) => this.#handle(chunk, controller),
flush: (controller) => this.#handle("\r\n", controller),
});
}
#handle(chunk, controller) {
chunk = this.#buf + chunk;
while (true) {
const lfIndex = chunk.indexOf("\n");
if (lfIndex !== -1) {
let crOrLfIndex = lfIndex;
if (chunk[lfIndex - 1] === "\r") {
crOrLfIndex--;
}
controller.enqueue(chunk.slice(0, crOrLfIndex));
chunk = chunk.slice(lfIndex + 1);
continue;
}
break;
}
this.#buf = chunk;
}
}
async function main () {
// Paths based on your examples:
const pathIn = 'example.txt';
const pathOut = `${pathIn}.final.txt`;
// Create file handles to the target file paths:
// See: https://nodejs.org/docs/latest-v18.x/api/fs.html#fspromisesopenpath-flags-mode
const fhIn = await open(pathIn);
// The "w" flag means: Open file for writing. The file is created (if it does not exist) or truncated (if it exists).
// See: https://nodejs.org/docs/latest-v18.x/api/fs.html#file-system-flags
const fhOut = await open(pathOut, 'w');
// Create a web-standard WritableStream from the output file handle:
// See: https://nodejs.org/docs/latest-v18.x/api/stream.html#streamwritabletowebstreamwritable
const writable = Writable.toWeb(fhOut.createWriteStream({encoding: 'utf8'}));
const writer = writable.getWriter();
// A function abstraction for writing a text chunk to the output file stream:
const write = (text) => writer.ready.then(() => writer.write(text));
// Crate a web-standard ReadableStream from the input file handle,
// then pipe through a text decoder and break/collect the emitted chunks into lines:
// See: https://nodejs.org/docs/latest-v18.x/api/fs.html#filehandlereadablewebstream
const readable = fhIn.readableWebStream()
.pipeThrough(new TextDecoderStream())
.pipeThrough(new TextLineStream());
for await (const line of readable) {
// Handle each text line in here:
// For example: get the length of each line,
// and if it's greater than 0, write it as a line to the output stream:
const {length} = line;
if (length > 0) await write(`${length}\n`);
}
}
main();
Here's the CLI output of using the program on an example text file with some Lorem ipsum lines:
% node --version
v18.12.1
% ls
example.txt module.mjs
% cat example.txt
lorem
ipsum
dolor
sit
amet
% node module.mjs
% cat example.txt.final.txt
5
5
5
3
4

Related

How can I optimize my JavaScript code to handle large log files (over 1 GB)? [duplicate]

I need to do some parsing of large (5-10 Gb)logfiles in Javascript/Node.js (I'm using Cube).
The logline looks something like:
10:00:43.343423 I'm a friendly log message. There are 5 cats, and 7 dogs. We are in state "SUCCESS".
We need to read each line, do some parsing (e.g. strip out 5, 7 and SUCCESS), then pump this data into Cube (https://github.com/square/cube) using their JS client.
Firstly, what is the canonical way in Node to read in a file, line by line?
It seems to be fairly common question online:
http://www.quora.com/What-is-the-best-way-to-read-a-file-line-by-line-in-node-js
Read a file one line at a time in node.js?
A lot of the answers seem to point to a bunch of third-party modules:
https://github.com/nickewing/line-reader
https://github.com/jahewson/node-byline
https://github.com/pkrumins/node-lazy
https://github.com/Gagle/Node-BufferedReader
However, this seems like a fairly basic task - surely, there's a simple way within the stdlib to read in a textfile, line-by-line?
Secondly, I then need to process each line (e.g. convert the timestamp into a Date object, and extract useful fields).
What's the best way to do this, maximising throughput? Is there some way that won't block on either reading in each line, or on sending it to Cube?
Thirdly - I'm guessing using string splits, and the JS equivalent of contains (IndexOf != -1?) will be a lot faster than regexes? Has anybody had much experience in parsing massive amounts of text data in Node.js?
I searched for a solution to parse very large files (gbs) line by line using a stream. All the third-party libraries and examples did not suit my needs since they processed the files not line by line (like 1 , 2 , 3 , 4 ..) or read the entire file to memory
The following solution can parse very large files, line by line using stream & pipe. For testing I used a 2.1 gb file with 17.000.000 records. Ram usage did not exceed 60 mb.
First, install the event-stream package:
npm install event-stream
Then:
var fs = require('fs')
, es = require('event-stream');
var lineNr = 0;
var s = fs.createReadStream('very-large-file.csv')
.pipe(es.split())
.pipe(es.mapSync(function(line){
// pause the readstream
s.pause();
lineNr += 1;
// process line here and call s.resume() when rdy
// function below was for logging memory usage
logMemoryUsage(lineNr);
// resume the readstream, possibly from a callback
s.resume();
})
.on('error', function(err){
console.log('Error while reading file.', err);
})
.on('end', function(){
console.log('Read entire file.')
})
);
Please let me know how it goes!
You can use the inbuilt readline package, see docs here. I use stream to create a new output stream.
var fs = require('fs'),
readline = require('readline'),
stream = require('stream');
var instream = fs.createReadStream('/path/to/file');
var outstream = new stream;
outstream.readable = true;
outstream.writable = true;
var rl = readline.createInterface({
input: instream,
output: outstream,
terminal: false
});
rl.on('line', function(line) {
console.log(line);
//Do your stuff ...
//Then write to output stream
rl.write(line);
});
Large files will take some time to process. Do tell if it works.
I really liked #gerard answer which is actually deserves to be the correct answer here. I made some improvements:
Code is in a class (modular)
Parsing is included
Ability to resume is given to the outside in case there is an asynchronous job is chained to reading the CSV like inserting to DB, or a HTTP request
Reading in chunks/batche sizes that
user can declare. I took care of encoding in the stream too, in case
you have files in different encoding.
Here's the code:
'use strict'
const fs = require('fs'),
util = require('util'),
stream = require('stream'),
es = require('event-stream'),
parse = require("csv-parse"),
iconv = require('iconv-lite');
class CSVReader {
constructor(filename, batchSize, columns) {
this.reader = fs.createReadStream(filename).pipe(iconv.decodeStream('utf8'))
this.batchSize = batchSize || 1000
this.lineNumber = 0
this.data = []
this.parseOptions = {delimiter: '\t', columns: true, escape: '/', relax: true}
}
read(callback) {
this.reader
.pipe(es.split())
.pipe(es.mapSync(line => {
++this.lineNumber
parse(line, this.parseOptions, (err, d) => {
this.data.push(d[0])
})
if (this.lineNumber % this.batchSize === 0) {
callback(this.data)
}
})
.on('error', function(){
console.log('Error while reading file.')
})
.on('end', function(){
console.log('Read entirefile.')
}))
}
continue () {
this.data = []
this.reader.resume()
}
}
module.exports = CSVReader
So basically, here is how you will use it:
let reader = CSVReader('path_to_file.csv')
reader.read(() => reader.continue())
I tested this with a 35GB CSV file and it worked for me and that's why I chose to build it on #gerard's answer, feedbacks are welcomed.
I used https://www.npmjs.com/package/line-by-line for reading more than 1 000 000 lines from a text file. In this case, an occupied capacity of RAM was about 50-60 megabyte.
const LineByLineReader = require('line-by-line'),
lr = new LineByLineReader('big_file.txt');
lr.on('error', function (err) {
// 'err' contains error object
});
lr.on('line', function (line) {
// pause emitting of lines...
lr.pause();
// ...do your asynchronous line processing..
setTimeout(function () {
// ...and continue emitting lines.
lr.resume();
}, 100);
});
lr.on('end', function () {
// All lines are read, file is closed now.
});
The Node.js Documentation offers a very elegant example using the Readline module.
Example: Read File Stream Line-by-Line
const { once } = require('node:events');
const fs = require('fs');
const readline = require('readline');
const rl = readline.createInterface({
input: fs.createReadStream('sample.txt'),
crlfDelay: Infinity
});
rl.on('line', (line) => {
console.log(`Line from file: ${line}`);
});
await once(rl, 'close');
Note: we use the crlfDelay option to recognize all instances of CR LF ('\r\n') as a single line break.
Apart from read the big file line by line, you also can read it chunk by chunk. For more refer to this article
var offset = 0;
var chunkSize = 2048;
var chunkBuffer = new Buffer(chunkSize);
var fp = fs.openSync('filepath', 'r');
var bytesRead = 0;
while(bytesRead = fs.readSync(fp, chunkBuffer, 0, chunkSize, offset)) {
offset += bytesRead;
var str = chunkBuffer.slice(0, bytesRead).toString();
var arr = str.split('\n');
if(bytesRead = chunkSize) {
// the last item of the arr may be not a full line, leave it to the next chunk
offset -= arr.pop().length;
}
lines.push(arr);
}
console.log(lines);
I had the same problem yet. After comparing several modules that seem to have this feature, I decided to do it myself, it's simpler than I thought.
gist: https://gist.github.com/deemstone/8279565
var fetchBlock = lineByline(filepath, onEnd);
fetchBlock(function(lines, start){ ... }); //lines{array} start{int} lines[0] No.
It cover the file opened in a closure, that fetchBlock() returned will fetch a block from the file, end split to array (will deal the segment from last fetch).
I've set the block size to 1024 for each read operation. This may have bugs, but code logic is obvious, try it yourself.
Reading / Writing files using stream with the native nodejs modules (fs, readline):
const fs = require('fs');
const readline = require('readline');
const rl = readline.createInterface({
input: fs.createReadStream('input.json'),
output: fs.createWriteStream('output.json')
});
rl.on('line', function(line) {
console.log(line);
// Do any 'line' processing if you want and then write to the output file
this.output.write(`${line}\n`);
});
rl.on('close', function() {
console.log(`Created "${this.output.path}"`);
});
Based on this questions answer I implemented a class you can use to read a file synchronously line-by-line with fs.readSync(). You can make this "pause" and "resume" by using a Q promise (jQuery seems to require a DOM so cant run it with nodejs):
var fs = require('fs');
var Q = require('q');
var lr = new LineReader(filenameToLoad);
lr.open();
var promise;
workOnLine = function () {
var line = lr.readNextLine();
promise = complexLineTransformation(line).then(
function() {console.log('ok');workOnLine();},
function() {console.log('error');}
);
}
workOnLine();
complexLineTransformation = function (line) {
var deferred = Q.defer();
// ... async call goes here, in callback: deferred.resolve('done ok'); or deferred.reject(new Error(error));
return deferred.promise;
}
function LineReader (filename) {
this.moreLinesAvailable = true;
this.fd = undefined;
this.bufferSize = 1024*1024;
this.buffer = new Buffer(this.bufferSize);
this.leftOver = '';
this.read = undefined;
this.idxStart = undefined;
this.idx = undefined;
this.lineNumber = 0;
this._bundleOfLines = [];
this.open = function() {
this.fd = fs.openSync(filename, 'r');
};
this.readNextLine = function () {
if (this._bundleOfLines.length === 0) {
this._readNextBundleOfLines();
}
this.lineNumber++;
var lineToReturn = this._bundleOfLines[0];
this._bundleOfLines.splice(0, 1); // remove first element (pos, howmany)
return lineToReturn;
};
this.getLineNumber = function() {
return this.lineNumber;
};
this._readNextBundleOfLines = function() {
var line = "";
while ((this.read = fs.readSync(this.fd, this.buffer, 0, this.bufferSize, null)) !== 0) { // read next bytes until end of file
this.leftOver += this.buffer.toString('utf8', 0, this.read); // append to leftOver
this.idxStart = 0
while ((this.idx = this.leftOver.indexOf("\n", this.idxStart)) !== -1) { // as long as there is a newline-char in leftOver
line = this.leftOver.substring(this.idxStart, this.idx);
this._bundleOfLines.push(line);
this.idxStart = this.idx + 1;
}
this.leftOver = this.leftOver.substring(this.idxStart);
if (line !== "") {
break;
}
}
};
}
node-byline uses streams, so i would prefer that one for your huge files.
for your date-conversions i would use moment.js.
for maximising your throughput you could think about using a software-cluster. there are some nice-modules which wrap the node-native cluster-module quite well. i like cluster-master from isaacs. e.g. you could create a cluster of x workers which all compute a file.
for benchmarking splits vs regexes use benchmark.js. i havent tested it until now. benchmark.js is available as a node-module
import * as csv from 'fast-csv';
import * as fs from 'fs';
interface Row {
[s: string]: string;
}
type RowCallBack = (data: Row, index: number) => object;
export class CSVReader {
protected file: string;
protected csvOptions = {
delimiter: ',',
headers: true,
ignoreEmpty: true,
trim: true
};
constructor(file: string, csvOptions = {}) {
if (!fs.existsSync(file)) {
throw new Error(`File ${file} not found.`);
}
this.file = file;
this.csvOptions = Object.assign({}, this.csvOptions, csvOptions);
}
public read(callback: RowCallBack): Promise < Array < object >> {
return new Promise < Array < object >> (resolve => {
const readStream = fs.createReadStream(this.file);
const results: Array < any > = [];
let index = 0;
const csvStream = csv.parse(this.csvOptions).on('data', async (data: Row) => {
index++;
results.push(await callback(data, index));
}).on('error', (err: Error) => {
console.error(err.message);
throw err;
}).on('end', () => {
resolve(results);
});
readStream.pipe(csvStream);
});
}
}
import { CSVReader } from '../src/helpers/CSVReader';
(async () => {
const reader = new CSVReader('./database/migrations/csv/users.csv');
const users = await reader.read(async data => {
return {
username: data.username,
name: data.name,
email: data.email,
cellPhone: data.cell_phone,
homePhone: data.home_phone,
roleId: data.role_id,
description: data.description,
state: data.state,
};
});
console.log(users);
})();
I have made a node module to read large file asynchronously text or JSON.
Tested on large files.
var fs = require('fs')
, util = require('util')
, stream = require('stream')
, es = require('event-stream');
module.exports = FileReader;
function FileReader(){
}
FileReader.prototype.read = function(pathToFile, callback){
var returnTxt = '';
var s = fs.createReadStream(pathToFile)
.pipe(es.split())
.pipe(es.mapSync(function(line){
// pause the readstream
s.pause();
//console.log('reading line: '+line);
returnTxt += line;
// resume the readstream, possibly from a callback
s.resume();
})
.on('error', function(){
console.log('Error while reading file.');
})
.on('end', function(){
console.log('Read entire file.');
callback(returnTxt);
})
);
};
FileReader.prototype.readJSON = function(pathToFile, callback){
try{
this.read(pathToFile, function(txt){callback(JSON.parse(txt));});
}
catch(err){
throw new Error('json file is not valid! '+err.stack);
}
};
Just save the file as file-reader.js, and use it like this:
var FileReader = require('./file-reader');
var fileReader = new FileReader();
fileReader.readJSON(__dirname + '/largeFile.json', function(jsonObj){/*callback logic here*/});

read text line by line from text file Node js [duplicate]

I am trying to read a large file one line at a time. I found a question on Quora that dealt with the subject but I'm missing some connections to make the whole thing fit together.
var Lazy=require("lazy");
new Lazy(process.stdin)
.lines
.forEach(
function(line) {
console.log(line.toString());
}
);
process.stdin.resume();
The bit that I'd like to figure out is how I might read one line at a time from a file instead of STDIN as in this sample.
I tried:
fs.open('./VeryBigFile.csv', 'r', '0666', Process);
function Process(err, fd) {
if (err) throw err;
// DO lazy read
}
but it's not working. I know that in a pinch I could fall back to using something like PHP, but I would like to figure this out.
I don't think the other answer would work as the file is much larger than the server I'm running it on has memory for.
Since Node.js v0.12 and as of Node.js v4.0.0, there is a stable readline core module. Here's the easiest way to read lines from a file, without any external modules:
const fs = require('fs');
const readline = require('readline');
async function processLineByLine() {
const fileStream = fs.createReadStream('input.txt');
const rl = readline.createInterface({
input: fileStream,
crlfDelay: Infinity
});
// Note: we use the crlfDelay option to recognize all instances of CR LF
// ('\r\n') in input.txt as a single line break.
for await (const line of rl) {
// Each line in input.txt will be successively available here as `line`.
console.log(`Line from file: ${line}`);
}
}
processLineByLine();
Or alternatively:
var lineReader = require('readline').createInterface({
input: require('fs').createReadStream('file.in')
});
lineReader.on('line', function (line) {
console.log('Line from file:', line);
});
The last line is read correctly (as of Node v0.12 or later), even if there is no final \n.
UPDATE: this example has been added to Node's API official documentation.
For such a simple operation there shouldn't be any dependency on third-party modules. Go easy.
var fs = require('fs'),
readline = require('readline');
var rd = readline.createInterface({
input: fs.createReadStream('/path/to/file'),
output: process.stdout,
console: false
});
rd.on('line', function(line) {
console.log(line);
});
Update in 2019
An awesome example is already posted on official Nodejs documentation. here
This requires the latest Nodejs is installed on your machine. >11.4
const fs = require('fs');
const readline = require('readline');
async function processLineByLine() {
const fileStream = fs.createReadStream('input.txt');
const rl = readline.createInterface({
input: fileStream,
crlfDelay: Infinity
});
// Note: we use the crlfDelay option to recognize all instances of CR LF
// ('\r\n') in input.txt as a single line break.
for await (const line of rl) {
// Each line in input.txt will be successively available here as `line`.
console.log(`Line from file: ${line}`);
}
}
processLineByLine();
You don't have to open the file, but instead, you have to create a ReadStream.
fs.createReadStream
Then pass that stream to Lazy
require('fs').readFileSync('file.txt', 'utf-8').split(/\r?\n/).forEach(function(line){
console.log(line);
})
there is a very nice module for reading a file line by line, it's called line-reader
with it you simply just write:
var lineReader = require('line-reader');
lineReader.eachLine('file.txt', function(line, last) {
console.log(line);
// do whatever you want with line...
if(last){
// or check if it's the last one
}
});
you can even iterate the file with a "java-style" interface, if you need more control:
lineReader.open('file.txt', function(reader) {
if (reader.hasNextLine()) {
reader.nextLine(function(line) {
console.log(line);
});
}
});
Old topic, but this works:
var rl = readline.createInterface({
input : fs.createReadStream('/path/file.txt'),
output: process.stdout,
terminal: false
})
rl.on('line',function(line){
console.log(line) //or parse line
})
Simple. No need for an external module.
You can always roll your own line reader. I have'nt benchmarked this snippet yet, but it correctly splits the incoming stream of chunks into lines without the trailing '\n'
var last = "";
process.stdin.on('data', function(chunk) {
var lines, i;
lines = (last+chunk).split("\n");
for(i = 0; i < lines.length - 1; i++) {
console.log("line: " + lines[i]);
}
last = lines[i];
});
process.stdin.on('end', function() {
console.log("line: " + last);
});
process.stdin.resume();
I did come up with this when working on a quick log parsing script that needed to accumulate data during the log parsing and I felt that it would nice to try doing this using js and node instead of using perl or bash.
Anyway, I do feel that small nodejs scripts should be self contained and not rely on third party modules so after reading all the answers to this question, each using various modules to handle line parsing, a 13 SLOC native nodejs solution might be of interest .
With the carrier module:
var carrier = require('carrier');
process.stdin.resume();
carrier.carry(process.stdin, function(line) {
console.log('got one line: ' + line);
});
I ended up with a massive, massive memory leak using Lazy to read line by line when trying to then process those lines and write them to another stream due to the way drain/pause/resume in node works (see: http://elegantcode.com/2011/04/06/taking-baby-steps-with-node-js-pumping-data-between-streams/ (i love this guy btw)). I haven't looked closely enough at Lazy to understand exactly why, but I couldn't pause my read stream to allow for a drain without Lazy exiting.
I wrote the code to process massive csv files into xml docs, you can see the code here: https://github.com/j03m/node-csv2xml
If you run the previous revisions with Lazy line it leaks. The latest revision doesn't leak at all and you can probably use it as the basis for a reader/processor. Though I have some custom stuff in there.
Edit: I guess I should also note that my code with Lazy worked fine until I found myself writing large enough xml fragments that drain/pause/resume because a necessity. For smaller chunks it was fine.
In most cases this should be enough:
const fs = require("fs")
fs.readFile('./file', 'utf-8', (err, file) => {
const lines = file.split('\n')
for (let line of lines)
console.log(line)
});
Edit:
Use a transform stream.
With a BufferedReader you can read lines.
new BufferedReader ("lorem ipsum", { encoding: "utf8" })
.on ("error", function (error){
console.log ("error: " + error);
})
.on ("line", function (line){
console.log ("line: " + line);
})
.on ("end", function (){
console.log ("EOF");
})
.read ();
I was frustrated by the lack of a comprehensive solution for this, so I put together my own attempt (git / npm). Copy-pasted list of features:
Interactive line processing (callback-based, no loading the entire file into RAM)
Optionally, return all lines in an array (detailed or raw mode)
Interactively interrupt streaming, or perform map/filter like processing
Detect any newline convention (PC/Mac/Linux)
Correct eof / last line treatment
Correct handling of multi-byte UTF-8 characters
Retrieve byte offset and byte length information on per-line basis
Random access, using line-based or byte-based offsets
Automatically map line-offset information, to speed up random access
Zero dependencies
Tests
NIH? You decide :-)
Since posting my original answer, I found that split is a very easy to use node module for line reading in a file; Which also accepts optional parameters.
var split = require('split');
fs.createReadStream(file)
.pipe(split())
.on('data', function (line) {
//each chunk now is a seperate line!
});
Haven't tested on very large files. Let us know if you do.
function createLineReader(fileName){
var EM = require("events").EventEmitter
var ev = new EM()
var stream = require("fs").createReadStream(fileName)
var remainder = null;
stream.on("data",function(data){
if(remainder != null){//append newly received data chunk
var tmp = new Buffer(remainder.length+data.length)
remainder.copy(tmp)
data.copy(tmp,remainder.length)
data = tmp;
}
var start = 0;
for(var i=0; i<data.length; i++){
if(data[i] == 10){ //\n new line
var line = data.slice(start,i)
ev.emit("line", line)
start = i+1;
}
}
if(start<data.length){
remainder = data.slice(start);
}else{
remainder = null;
}
})
stream.on("end",function(){
if(null!=remainder) ev.emit("line",remainder)
})
return ev
}
//---------main---------------
fileName = process.argv[2]
lineReader = createLineReader(fileName)
lineReader.on("line",function(line){
console.log(line.toString())
//console.log("++++++++++++++++++++")
})
I wanted to tackle this same problem, basically what in Perl would be:
while (<>) {
process_line($_);
}
My use case was just a standalone script, not a server, so synchronous was fine. These were my criteria:
The minimal synchronous code that could reuse in many projects.
No limits on file size or number of lines.
No limits on length of lines.
Able to handle full Unicode in UTF-8, including characters beyond the BMP.
Able to handle *nix and Windows line endings (old-style Mac not needed for me).
Line endings character(s) to be included in lines.
Able to handle last line with or without end-of-line characters.
Not use any external libraries not included in the node.js distribution.
This is a project for me to get a feel for low-level scripting type code in node.js and decide how viable it is as a replacement for other scripting languages like Perl.
After a surprising amount of effort and a couple of false starts this is the code I came up with. It's pretty fast but less trivial than I would've expected: (fork it on GitHub)
var fs = require('fs'),
StringDecoder = require('string_decoder').StringDecoder,
util = require('util');
function lineByLine(fd) {
var blob = '';
var blobStart = 0;
var blobEnd = 0;
var decoder = new StringDecoder('utf8');
var CHUNK_SIZE = 16384;
var chunk = new Buffer(CHUNK_SIZE);
var eolPos = -1;
var lastChunk = false;
var moreLines = true;
var readMore = true;
// each line
while (moreLines) {
readMore = true;
// append more chunks from the file onto the end of our blob of text until we have an EOL or EOF
while (readMore) {
// do we have a whole line? (with LF)
eolPos = blob.indexOf('\n', blobStart);
if (eolPos !== -1) {
blobEnd = eolPos;
readMore = false;
// do we have the last line? (no LF)
} else if (lastChunk) {
blobEnd = blob.length;
readMore = false;
// otherwise read more
} else {
var bytesRead = fs.readSync(fd, chunk, 0, CHUNK_SIZE, null);
lastChunk = bytesRead !== CHUNK_SIZE;
blob += decoder.write(chunk.slice(0, bytesRead));
}
}
if (blobStart < blob.length) {
processLine(blob.substring(blobStart, blobEnd + 1));
blobStart = blobEnd + 1;
if (blobStart >= CHUNK_SIZE) {
// blobStart is in characters, CHUNK_SIZE is in octets
var freeable = blobStart / CHUNK_SIZE;
// keep blob from growing indefinitely, not as deterministic as I'd like
blob = blob.substring(CHUNK_SIZE);
blobStart -= CHUNK_SIZE;
blobEnd -= CHUNK_SIZE;
}
} else {
moreLines = false;
}
}
}
It could probably be cleaned up further, it was the result of trial and error.
Generator based line reader: https://github.com/neurosnap/gen-readlines
var fs = require('fs');
var readlines = require('gen-readlines');
fs.open('./file.txt', 'r', function(err, fd) {
if (err) throw err;
fs.fstat(fd, function(err, stats) {
if (err) throw err;
for (var line of readlines(fd, stats.size)) {
console.log(line.toString());
}
});
});
A new function was added in Node.js v18.11.0 to read files line by line
filehandle.readLines([options])
This is how you use this with a text file you want to read
import { open } from 'node:fs/promises';
myFileReader();
async function myFileReader() {
const file = await open('./TextFileName.txt');
for await (const line of file.readLines()) {
console.log(line)
}
}
To understand more read Node.js documentation here is the link for file system readlines():
https://nodejs.org/api/fs.html#filehandlereadlinesoptions
If you want to read a file line by line and writing this in another:
var fs = require('fs');
var readline = require('readline');
var Stream = require('stream');
function readFileLineByLine(inputFile, outputFile) {
var instream = fs.createReadStream(inputFile);
var outstream = new Stream();
outstream.readable = true;
outstream.writable = true;
var rl = readline.createInterface({
input: instream,
output: outstream,
terminal: false
});
rl.on('line', function (line) {
fs.appendFileSync(outputFile, line + '\n');
});
};
var fs = require('fs');
function readfile(name,online,onend,encoding) {
var bufsize = 1024;
var buffer = new Buffer(bufsize);
var bufread = 0;
var fd = fs.openSync(name,'r');
var position = 0;
var eof = false;
var data = "";
var lines = 0;
encoding = encoding || "utf8";
function readbuf() {
bufread = fs.readSync(fd,buffer,0,bufsize,position);
position += bufread;
eof = bufread ? false : true;
data += buffer.toString(encoding,0,bufread);
}
function getLine() {
var nl = data.indexOf("\r"), hasnl = nl !== -1;
if (!hasnl && eof) return fs.closeSync(fd), online(data,++lines), onend(lines);
if (!hasnl && !eof) readbuf(), nl = data.indexOf("\r"), hasnl = nl !== -1;
if (!hasnl) return process.nextTick(getLine);
var line = data.substr(0,nl);
data = data.substr(nl+1);
if (data[0] === "\n") data = data.substr(1);
online(line,++lines);
process.nextTick(getLine);
}
getLine();
}
I had the same problem and came up with above solution
looks simular to others but is aSync and can read large files very quickly
Hopes this helps
Two questions we must ask ourselves while doing such operations are:
What's the amount of memory used to perform it?
Is the memory consumption increasing drastically with the file size?
Solutions like require('fs').readFileSync() loads the whole file into memory. That means that the amount of memory required to perform operations will be almost equivalent to the file size. We should avoid these for anything larger than 50mbs
We can easily track the amount of memory used by a function by placing these lines of code after the function invocation :
const used = process.memoryUsage().heapUsed / 1024 / 1024;
console.log(
`The script uses approximately ${Math.round(used * 100) / 100} MB`
);
Right now the best way to read particular lines from a large file is using node's readline. The documentation has amazing examples.
This is my favorite way of going through a file, a simple native solution for a progressive (as in not a "slurp" or all-in-memory way) file read with modern async/await. It's a solution that I find "natural" when processing large text files without having to resort to the readline package or any non-core dependency.
let buf = '';
for await ( const chunk of fs.createReadStream('myfile') ) {
const lines = buf.concat(chunk).split(/\r?\n/);
buf = lines.pop() ?? '';
for( const line of lines ) {
console.log(line);
}
}
if(buf.length) console.log(buf); // last line, if file does not end with newline
You can adjust encoding in the fs.createReadStream or use chunk.toString(<arg>). Also this let's you better fine-tune the line splitting to your taste, ie. use .split(/\n+/) to skip empty lines and control the chunk size with fs.createReadStream('myfile', { highWaterMark: <chunkSize> }).
Don't forget to create a function like processLine(line) to avoid repeating the line processing code twice due to the ending buf leftover. Unfortunately, the ReadStream instance does not update its end-of-file flags in this setup, so there's no way, afaik, to detect within the loop that we're in the last iteration without some more verbose tricks like comparing the file size from a fs.Stats() with .bytesRead. Hence the final buf processing solution, unless you're absolutely sure your file ends with a newline \n, in which case the for await loop should suffice.
Performance Considerations
Chunk sizes are important for performance, the default is 64k for text files and, for multi MB files, larger chunks can improve speed by an order of magnitude.
The above snippet runs at least the same speed (or even 5% faster sometimes) as code based on NodeJS v18's fs.readLine() or based on the readline module (the accepted answer), once you tune highWaterMark to something that your machine can handle, ie. setting it to the same size as the file, if your available memory allows it, is the fastest.
In any case, any of NodeJS line-reading answers here are an order of magnitude slower than the Perl or native *Nix solutions.
Similar alternatives
★ If you prefer the evented asynchronous version, this would be it:
let buf = '';
fs.createReadStream('myfile')
.on('data', chunk => {
const lines = buf.concat(chunk).split(/\r?\n/);
buf = lines.pop();
for( const line of lines ) {
console.log(line);
}
})
.on('end', () => buf.length && console.log(buf) );
★ Now if you don't mind importing the stream core package, then this is the equivalent piped stream version, which allows for chaining transforms like gzip decompression:
const { Writable } = require('stream');
let buf = '';
fs.createReadStream('myfile').pipe(
new Writable({
write: (chunk, enc, next) => {
const lines = buf.concat(chunk).split(/\r?\n/);
buf = lines.pop();
for (const line of lines) {
console.log(line);
}
next();
}
})
).on('finish', () => buf.length && console.log(buf) );
I have a little module which does this well and is used by quite a few other projects npm readline Note thay in node v10 there is a native readline module so I republished my module as linebyline https://www.npmjs.com/package/linebyline
if you dont want to use the module the function is very simple:
var fs = require('fs'),
EventEmitter = require('events').EventEmitter,
util = require('util'),
newlines = [
13, // \r
10 // \n
];
var readLine = module.exports = function(file, opts) {
if (!(this instanceof readLine)) return new readLine(file);
EventEmitter.call(this);
opts = opts || {};
var self = this,
line = [],
lineCount = 0,
emit = function(line, count) {
self.emit('line', new Buffer(line).toString(), count);
};
this.input = fs.createReadStream(file);
this.input.on('open', function(fd) {
self.emit('open', fd);
})
.on('data', function(data) {
for (var i = 0; i < data.length; i++) {
if (0 <= newlines.indexOf(data[i])) { // Newline char was found.
lineCount++;
if (line.length) emit(line, lineCount);
line = []; // Empty buffer.
} else {
line.push(data[i]); // Buffer new line data.
}
}
}).on('error', function(err) {
self.emit('error', err);
}).on('end', function() {
// Emit last line if anything left over since EOF won't trigger it.
if (line.length){
lineCount++;
emit(line, lineCount);
}
self.emit('end');
}).on('close', function() {
self.emit('close');
});
};
util.inherits(readLine, EventEmitter);
Another solution is to run logic via sequential executor nsynjs. It reads file line-by-line using node readline module, and it doesn't use promises or recursion, therefore not going to fail on large files. Here is how the code will looks like:
var nsynjs = require('nsynjs');
var textFile = require('./wrappers/nodeReadline').textFile; // this file is part of nsynjs
function process(textFile) {
var fh = new textFile();
fh.open('path/to/file');
var s;
while (typeof(s = fh.readLine(nsynjsCtx).data) != 'undefined')
console.log(s);
fh.close();
}
var ctx = nsynjs.run(process,{},textFile,function () {
console.log('done');
});
Code above is based on this exampe: https://github.com/amaksr/nsynjs/blob/master/examples/node-readline/index.js
i use this:
function emitLines(stream, re){
re = re && /\n/;
var buffer = '';
stream.on('data', stream_data);
stream.on('end', stream_end);
function stream_data(data){
buffer += data;
flush();
}//stream_data
function stream_end(){
if(buffer) stream.emmit('line', buffer);
}//stream_end
function flush(){
var re = /\n/;
var match;
while(match = re.exec(buffer)){
var index = match.index + match[0].length;
stream.emit('line', buffer.substring(0, index));
buffer = buffer.substring(index);
re.lastIndex = 0;
}
}//flush
}//emitLines
use this function on a stream and listen to the line events that is will emit.
gr-
While you should probably use the readline module as the top answer suggests, readline appears to be oriented toward command line interfaces rather than line reading. It's also a little bit more opaque regarding buffering. (Anyone who needs a streaming line oriented reader probably will want to tweak buffer sizes). The readline module is ~1000 lines while this, with stats and tests, is 34.
const EventEmitter = require('events').EventEmitter;
class LineReader extends EventEmitter{
constructor(f, delim='\n'){
super();
this.totalChars = 0;
this.totalLines = 0;
this.leftover = '';
f.on('data', (chunk)=>{
this.totalChars += chunk.length;
let lines = chunk.split(delim);
if (lines.length === 1){
this.leftover += chunk;
return;
}
lines[0] = this.leftover + lines[0];
this.leftover = lines[lines.length-1];
if (this.leftover) lines.pop();
this.totalLines += lines.length;
for (let l of lines) this.onLine(l);
});
// f.on('error', ()=>{});
f.on('end', ()=>{console.log('chars', this.totalChars, 'lines', this.totalLines)});
}
onLine(l){
this.emit('line', l);
}
}
//Command line test
const f = require('fs').createReadStream(process.argv[2], 'utf8');
const delim = process.argv[3];
const lineReader = new LineReader(f, delim);
lineReader.on('line', (line)=> console.log(line));
Here's an even shorter version, without the stats, at 19 lines:
class LineReader extends require('events').EventEmitter{
constructor(f, delim='\n'){
super();
this.leftover = '';
f.on('data', (chunk)=>{
let lines = chunk.split(delim);
if (lines.length === 1){
this.leftover += chunk;
return;
}
lines[0] = this.leftover + lines[0];
this.leftover = lines[lines.length-1];
if (this.leftover)
lines.pop();
for (let l of lines)
this.emit('line', l);
});
}
}
const fs = require("fs")
fs.readFile('./file', 'utf-8', (err, data) => {
var innerContent;
console.log("Asynchronous read: " + data.toString());
const lines = data.toString().split('\n')
for (let line of lines)
innerContent += line + '<br>';
});
I wrap the whole logic of daily line processing as a npm module: line-kit
https://www.npmjs.com/package/line-kit
// example
var count = 0
require('line-kit')(require('fs').createReadStream('/etc/issue'),
(line) => { count++; },
() => {console.log(`seen ${count} lines`)})
I use below code the read lines after verify that its not a directory and its not included in the list of files need not to be check.
(function () {
var fs = require('fs');
var glob = require('glob-fs')();
var path = require('path');
var result = 0;
var exclude = ['LICENSE',
path.join('e2e', 'util', 'db-ca', 'someother-file'),
path.join('src', 'favicon.ico')];
var files = [];
files = glob.readdirSync('**');
var allFiles = [];
var patternString = [
'trade',
'order',
'market',
'securities'
];
files.map((file) => {
try {
if (!fs.lstatSync(file).isDirectory() && exclude.indexOf(file) === -1) {
fs.readFileSync(file).toString().split(/\r?\n/).forEach(function(line){
patternString.map((pattern) => {
if (line.indexOf(pattern) !== -1) {
console.log(file + ' contain `' + pattern + '` in in line "' + line +'";');
result = 1;
}
});
});
}
} catch (e) {
console.log('Error:', e.stack);
}
});
process.exit(result);
})();
I have looked through all above answers, all of them use third-party library to solve it. It's have a simple solution in Node's API. e.g
const fs= require('fs')
let stream = fs.createReadStream('<filename>', { autoClose: true })
stream.on('data', chunk => {
let row = chunk.toString('ascii')
}))

NodeJS: Using Pipe To Write A File From A Readable Stream Gives Heap Memory Error

I am trying to create 150 million lines of data and write the data into a csv file so that I can insert the data into different databases with little modification.
I am using a few functions to generate seemingly random data and pushing the data into the writable stream.
The code that I have right now is unsuccessful at handling memory issue.
After a few hours of research, I am starting to think that I should not be pushing each data at the end of the for loop because it seems that the pipe method simply cannot handle garbage collection this way.
Also, I found a few StackOverFlow answers and NodeJS docs that recommend against using push at all.
However, I am very new to NodeJS and I feel like I am blocked and do not know how to proceed from here.
If someone can provide me any guidance on how to proceed and give me an example, I would really appreciate it.
Below is a part of my code to give you a better understanding of what I am trying to achieve.
P.S. -
I have found a way to write successfully handle memory issue without using pipe method at all --I used the drain event-- but I had to start from scratch and now I am curious to know if there is a simple way to handle this memory issue without completely changing this bit of code.
Also, I have been trying to avoid using any library because I feel like there should be a relatively easy tweak to make this work without using a library but please tell me if I am wrong. Thank you in advance.
// This is my target number of data
const targetDataNum = 150000000;
// Create readable stream
const readableStream = new Stream.Readable({
read() {}
});
// Create writable stream
const writableStream = fs.createWriteStream('./database/RDBMS/test.csv');
// Write columns first
writableStream.write('id, body, date, dp\n', 'utf8');
// Then, push a number of data to the readable stream (150M in this case)
for (var i = 1; i <= targetDataNum; i += 1) {
const id = i;
const body = lorem.paragraph(1);
const date = randomDate(new Date(2014, 0, 1), new Date());
const dp = randomNumber(1, 1000);
const data = `${id},${body},${date},${dp}\n`;
readableStream.push(data, 'utf8');
};
// Pipe readable stream to writeable stream
readableStream.pipe(writableStream);
// End the stream
readableStream.push(null);
Since you're new to streams, maybe start with an easier abstraction: generators. Generators generate data only when it is consumed (just like Streams should), but they don't have buffering and complicated constructors and methods.
This is just your for loop, moved into a generator function:
function * generateData(targetDataNum) {
for (var i = 1; i <= targetDataNum; i += 1) {
const id = i;
const body = lorem.paragraph(1);
const date = randomDate(new Date(2014, 0, 1), new Date());
const dp = randomNumber(1, 1000);
yield `${id},${body},${date},${dp}\n`;
}
}
In Node 12, you can create a Readable stream directly from any iterable, including generators and async generators:
const stream = Readable.from(generateData(), {encoding: 'utf8'})
stream.pipe(writableStream)
i suggest to try a solution like the following:
const { Readable } = require('readable-stream');
class CustomReadable extends Readable {
constructor(max, options = {}) {
super(options);
this.targetDataNum = max;
this.i = 1;
}
_read(size) {
if (i <= this.targetDataNum) {
// your code to build the csv content
this.push(data, 'utf8');
return;
}
this.push(null);
}
}
const rs = new CustomReadable(150000000);
rs.pipe(ws);
Just complete it with your portion of code to fill the csv and create the writable stream.
With this solution you leave calling the rs.push method to the internal _read stream method invoked until this.push(null) is not called. Probably before you were filling the internal stream buffer too fast calling push manually in a loop getting the out memory error.
Try pipeing to the WritableStream before you start pumping data into the ReadableStream and yield before you write the next chunk.
...
// Write columns first
writableStream.write('id, body, date, dp\n', 'utf8');
// Pipe readable stream to writeable stream
readableStream.pipe(writableStream);
// Then, push a number of data to the readable stream (150M in this case)
for (var i = 1; i <= targetDataNum; i += 1) {
const id = i;
const body = lorem.paragraph(1);
const date = randomDate(new Date(2014, 0, 1), new Date());
const dp = randomNumber(1, 1000);
const data = `${id},${body},${date},${dp}\n`;
readableStream.push(data, 'utf8');
// somehow YIELD for the STREAM to drain out.
};
...
The entire Stream implementation of Node.js relies on the fact that the wire is slow and that the CPU can actually have a downtime before the next chunk of data comes in from the stream source or till the next chunk of data has been written to the stream destination.
In the current implementation, since the for-loop has booked up the CPU, there is no downtime for the actual pipeing of the data to the writestream. You will be able to catch this if you watch cat test.csv which will not change while the loop is running.
As (I am sure) you know, pipe helps in guaranteeing that the data you are working with is buffered in memory only in chunks and not as a whole. But that guarantee only holds true if the CPU gets enough downtime to actually drain the data.
Having said all that, I wrapped your entire code into an async IIFE and ran it with an await for a setTimeout which ensures that I yield for the stream to drain the data.
let fs = require('fs');
let Stream = require('stream');
(async function () {
// This is my target number of data
const targetDataNum = 150000000;
// Create readable stream
const readableStream = new Stream.Readable({
read() { }
});
// Create writable stream
const writableStream = fs.createWriteStream('./test.csv');
// Write columns first
writableStream.write('id, body, date, dp\n', 'utf8');
// Pipe readable stream to writeable stream
readableStream.pipe(writableStream);
// Then, push a number of data to the readable stream (150M in this case)
for (var i = 1; i <= targetDataNum; i += 1) {
console.log(`Pushing ${i}`);
const id = i;
const body = `body${i}`;
const date = `date${i}`;
const dp = `dp${i}`;
const data = `${id},${body},${date},${dp}\n`;
readableStream.push(data, 'utf8');
await new Promise(resolve => setImmediate(resolve));
};
// End the stream
readableStream.push(null);
})();
This is what top looks like pretty much the whole time I am running this.
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
15213 binaek ** ** ****** ***** ***** * ***.* 0.5 *:**.** node
Notice the %MEM which stays more-or-less static.
You were running out of memory because you were pre-generating all the data in memory before you wrote any of it to disk. Instead, you need a strategy to write is as you generate so you don't have to hold large amounts of data in memory.
It does not seem like you need .pipe() here because you control the generation of the data (it's not coming from some random readStream).
So, you can just generate the data and immediately write it and handle the drain event when needed. Here's a runnable example (this creates a very large file):
const {once} = require('events');
const fs = require('fs');
// This is my target number of data
const targetDataNum = 150000000;
async function run() {
// Create writable stream
const writableStream = fs.createWriteStream('./test.csv');
// Write columns first
writableStream.write('id, body, date, dp\n', 'utf8');
// Then, push a number of data to the readable stream (150M in this case)
for (let i = 1; i <= targetDataNum; i += 1) {
const id = i;
const body = lorem.paragraph(1);
const date = randomDate(new Date(2014, 0, 1), new Date());
const dp = randomNumber(1, 1000);
const data = `${id},${body},${date},${dp}\n`;
const canWriteMore = writableStream.write(data);
if (!canWriteMore) {
// wait for stream to be ready for more writing
await once(writableStream, "drain");
}
}
writableStream.end();
}
run().then(() => {
console.log(done);
}).catch(err => {
console.log("got rejection: ", err);
});
// placeholders for the functions that were being used
function randomDate(low, high) {
let rand = randomNumber(low.getTime(), high.getTime());
return new Date(rand);
}
function randomNumber(low, high) {
return Math.floor(Math.random() * (high - low)) + low;
}
const lorem = {
paragraph: function() {
return "random paragraph";
}
}

Nodejs - removing substring from a huge file

I need to remove a substring (that appears only in specific known lines of the file) from a file.
there are simple solutions of reading all file data to a string, removing the substring, and then write the fixed data to the file.
here is a code I found in here:
Node js - Remove string from text file
var data = fs.readFileSync('banlist.txt', 'utf-8');
var newValue = data.replace(new RegEx("STRING_TO_REMOVE"), '');
fs.writeFileSync('banlist.txt', newValue, 'utf-8');
My problem is, that the file is huge - up to billion lines of logs, so I can't read all content to the memory.
Why not a simple transform stream and replace()? replace can take a callback as second parameter i.e. .replace(/bad1|bad2|bad3/g, filterWords) in case you need to replace words rather than remove them completely.
const fs = require("fs")
const { pipeline, Transform } = require("stream")
const { join } = require("path")
const readFile = fs.createReadStream("./words.txt")
const writeFile = fs.createWriteStream(
join(__dirname, "words-filtered.txt"),
"utf8"
)
const transformFile = new Transform({
transform(chunk, enc, next) {
let c = chunk.toString().replace(/bad/g, "replaced")
this.push(c)
next()
},
})
pipeline(readFile, transformFile, writeFile, (err) => {
if (err) {
console.log(err.message)
}
})
https://nodejs.org/api/fs.html#fs_fs_read_fd_buffer_offset_length_position_callback
Dont read the whole file at once... read a small buffered piece of it.. and look for your input with that buffered piece.... then increment your buffer starting position and do it again.... would recommend having each buffer start not at the end of the previous buffer... but overlap by at least the expected size of the data being sought so that you dont run into half of your data being at end of one buffer and other half at beginning of the other
You could use a file read stream. However, you would have to find a way to detect if the read data only contains part of the result.
What you probably want to do is use streams so that you are writing after partial reads. this example could probably work for you. you need to copy over the output text file ".tmp" over the original to get the same behavior in your question. It works by reading a chunk and then looking to see if you've come across a new line. then it processes that line, writes it, then removes it from the buffer. This should help with your memory problem.
var fs = require("fs");
var readStream = fs.createReadStream("./BFFile.txt", { encoding: "utf-8" });
var writeStream = fs.createWriteStream("./BFFile.txt.tmp");
const STRING_TO_REMOVE = "badword";
var buffer = ""
readStream.on("data", (chunk) => {
buffer += chunk;
var indexOfNewLine = buffer.search("\n");
while (indexOfNewLine !== -1) {
var line = buffer.substring(0, indexOfNewLine + 1);
buffer = buffer.substring(indexOfNewLine + 1, buffer.length);
line = line.replace(new RegExp(STRING_TO_REMOVE), "");
writeStream.write(line);
indexOfNewLine = buffer.search("\n");
}
})
readStream.on("end", () => {
buffer = buffer.replace(new RegExp(STRING_TO_REMOVE), "");
writeStream.write(buffer);
writeStream.close();
})
There are a few assumptions with this solution such as the data being UTF-8, there only being 1 bad word potentially per line, every line having some text (I didn't test for that), and that every line ends with new line and not some other line ending.
Heres the docs for streams in Node
another thought I had was to use pipe and a transform stream but that seems like over kill.
You can use this code to do it. I'm using fs stream. it's created for read huge files in small memory by chunks. docs
const fs = require('fs');
const readStream = fs.createReadStream('./XXXXX');
const writeStream = fs.createWriteStream('./XXXXXXX');
readStream.on('data', (chunk) => {
const data = chunk.toString().replace('STRING_TO_REMOVE', 'XXXXXX');
writeStream.write(data);
});
readStream.on('end', () => {
writeStream.close();
});

fs.createWriteStream doesn't use back-pressure when writing data to a file, causing high memory usage

Problem
I'm trying to scan a drive directory (recursively walk all the paths) and write all the paths to a file (as it's finding them) using fs.createWriteStream in order to keep the memory usage low, but it doesn't work, the memory usage reaches 2GB during the scan.
Expected
I was expecting fs.createWriteStream to automatically handle memory/disk usage at all times, keeping memory usage at a minimum with back-pressure.
Code
const fs = require('fs')
const walkdir = require('walkdir')
let dir = 'C:/'
let options = {
"max_depth": 0,
"track_inodes": true,
"return_object": false,
"no_return": true,
}
const wstream = fs.createWriteStream("C:/Users/USERNAME/Desktop/paths.txt")
let walker = walkdir(dir, options)
walker.on('path', (path) => {
wstream.write(path + '\n')
})
walker.on('end', (path) => {
wstream.end()
})
Is it because I'm not using .pipe()? I tried creating a new Stream.Readable({read{}}) and then inside the .on('path' emitter pushing paths into it with readable.push(path) but that didn't really work.
UPDATE:
Method 2:
I tried the proposed in the answers drain method but it doesn't help much, it does reduce memory usage to 500mb (which is still too much for a stream) but it slows down the code significantly (from seconds to minutes)
Method 3:
I also tried using readdirp, it uses even less memory (~400mb) and is faster but I don't know how to pause it and use the drain method there to reduce the memory usage further:
const readdirp = require('readdirp')
let dir = 'C:/'
const wstream = fs.createWriteStream("C:/Users/USERNAME/Desktop/paths.txt")
readdirp(dir, {alwaysStat: false, type: 'files_directories'})
.on('data', (entry) => {
wstream.write(`${entry.fullPath}\n`)
})
Method 4:
I also tried doing this operation with a custom recursive walker, and even though it uses only 30mb of memory, which is what I wanted, but it is like 10 times slower than the readdirp method and it is synchronous which is undesirable:
const fs = require('fs')
const path = require('path')
let dir = 'C:/'
function customRecursiveWalker(dir) {
fs.readdirSync(dir).forEach(file => {
let fullPath = path.join(dir, file)
// Folders
if (fs.lstatSync(fullPath).isDirectory()) {
fs.appendFileSync("C:/Users/USERNAME/Desktop/paths.txt", `${fullPath}\n`)
customRecursiveWalker(fullPath)
}
// Files
else {
fs.appendFileSync("C:/Users/USERNAME/Desktop/paths.txt", `${fullPath}\n`)
}
})
}
customRecursiveWalker(dir)
Preliminary observation: you've attempted to get the results you want using multiple approaches. One complication when comparing the approaches you used is that they do not all do the same work. If you run tests on file tree that contains only regular files, that tree does not contain mount points, you can probably compare the approaches fairly, but when you start adding mount points, symbolic links, etc, you may get different memory and time statistics merely due to the fact that one approach excludes files that another approach includes.
I've initially attempted a solution using readdirp, but unfortunately, but that library appears buggy to me. Running it on my system here, I got inconsistent results. One run would output 10Mb of data, another run with the same input parameters would output 22Mb, then I'd get another number, etc. I looked at the code and found that it does not respect the return value of push:
_push(entry) {
if (this.readable) {
this.push(entry);
}
}
As per the documentation the push method may return a false value, in which case the Readable stream should stop producing data and wait until _read is called again. readdirp entirely ignores that part of the specification. It is crucial to pay attention to the return value of push to get proper handling of back-pressure. There are also other things that seemed questionable in that code.
So I abandoned that and worked on a proof of concept showing how it could be done. The crucial parts are:
When the push method returns false it is imperative to stop adding data to the stream. Instead, we record where we were, and stop.
We start again only when _read is called.
If you uncomment the console.log statements that print START and STOP. You'll see them printed out in succession on the console. We start, produce data until Node tells us to stop, and then we stop, until Node tells us to start again, and so on.
const stream = require("stream");
const fs = require("fs");
const { readdir, lstat } = fs.promises;
const path = require("path");
class Walk extends stream.Readable {
constructor(root, maxDepth = Infinity) {
super();
this._maxDepth = maxDepth;
// These fields allow us to remember where we were when we have to pause our
// work.
// The path of the directory to process when we resume processing, and the
// depth of this directory.
this._curdir = [root, 1];
// The directories still to process.
this._dirs = [this._curdir];
// The list of files to process when we resume processing.
this._files = [];
// The location in `this._files` were to continue processing when we resume.
this._ix = 0;
// A flag recording whether or not the fetching of files is currently going
// on.
this._started = false;
}
async _fetch() {
// Recall where we were by loading the state in local variables.
let files = this._files;
let dirs = this._dirs;
let [dir, depth] = this._curdir;
let ix = this._ix;
while (true) {
// If we've gone past the end of the files we were processing, then
// just forget about them. This simplifies the code that follows a bit.
if (ix >= files.length) {
ix = 0;
files = [];
}
// Read directories until we have files to process.
while (!files.length) {
// We've read everything, end the stream.
if (dirs.length === 0) {
// This is how the stream API requires us to indicate the stream has
// ended.
this.push(null);
// We're no longer running.
this._started = false;
return;
}
// Here, we get the next directory to process and get the list of
// files in it.
[dir, depth] = dirs.pop();
try {
files = await readdir(dir, { withFileTypes: true });
}
catch (ex) {
// This is a proof-of-concept. In a real application, you should
// determine what exceptions you want to ignore (e.g. EPERM).
}
}
// Process each file.
for (; ix < files.length; ++ix) {
const dirent = files[ix];
// Don't include in the results those files that are not directories,
// files or symbolic links.
if (!(dirent.isFile() || dirent.isDirectory() || dirent.isSymbolicLink())) {
continue;
}
const fullPath = path.join(dir, dirent.name);
if (dirent.isDirectory() & depth < this._maxDepth) {
// Keep track that we need to walk this directory.
dirs.push([fullPath, depth + 1]);
}
// Finally, we can put the data into the stream!
if (!this.push(`${fullPath}\n`)) {
// If the push returned false, we have to stop pushing results to the
// stream until _read is called again, so we have to stop.
// Uncomment this if you want to see when the stream stops.
// console.log("STOP");
// Record where we were in our processing.
this._files = files;
// The element at ix *has* been processed, so ix + 1.
this._ix = ix + 1;
this._curdir = [dir, depth];
// We're stopping, so indicate that!
this._started = false;
return;
}
}
}
}
async _read() {
// Do not start the process that puts data on the stream over and over
// again.
if (this._started) {
return;
}
this._started = true; // Yep, we've started.
// Uncomment this if you want to see when the stream starts.
// console.log("START");
await this._fetch();
}
}
// Change the paths to something that makes sense for you.
stream.pipeline(new Walk("/home/", 5),
fs.createWriteStream("/tmp/paths3.txt"),
(err) => console.log("ended with", err));
When I run the first attempt you made with walkdir here, I get the following statistics:
Elapsed time (wall clock): 59 sec
Maximum resident set size: 2.90 GB
When I use the code I've shown above:
Elapsed time (wall clock): 35 sec
Maximum resident set size: 0.1 GB
The file tree I use for the tests produces a file listing of 792 MB
You could exploit the returned value from WritableStream.write(): it essentially states if you should continue to read or not. a WritableStream has an internal property that stores the threshold after which the buffer should be processed by the OS. The drain event will be emitted when the buffer has been flushed, i.e. you can call safely call WritableStream.write() without risking to excessively fill the buffer (which means the RAM). Luckily for you, walkdir let you control the process: you can emit pause(pause the walk. no more events will be emitted until resume) and resume(resume the walk) event from the walkdir object, pausing and resuming the writing process on you stream accordingly. Try with this:
let is_emitter_paused = false;
wstream.on('drain', (evt) => {
if (is_emitter_paused) {
walkdir.resume();
}
});
walkdir.on('path', function(path, stat) {
is_emitter_paused = !wstream.write(path + '\n');
if (is_emitter_paused) {
walkdir.pause();
}
});
Here's an implementation inspired by #Louis's answer. I think it's a bit easier to follow and in my minimal testing it performs about the same.
const fs = require('fs');
const path = require('path');
const stream = require('stream');
class Walker extends stream.Readable {
constructor(root = process.cwd(), maxDepth = Infinity) {
super();
// Dirs to process
this._dirs = [{ path: root, depth: 0 }];
// Max traversal depth
this._maxDepth = maxDepth;
// Files to flush
this._files = [];
}
_drain() {
while (this._files.length > 0) {
const file = this._files.pop();
if (file.isFile() || file.isDirectory() || file.isSymbolicLink()) {
const filePath = path.join(this._dir.path, file.name);
if (file.isDirectory() && this._maxDepth > this._dir.depth) {
// Add directory to be walked at a later time
this._dirs.push({ path: filePath, depth: this._dir.depth + 1 });
}
if (!this.push(`${filePath}\n`)) {
// Hault walking
return false;
}
}
}
if (this._dirs.length === 0) {
// Walking complete
this.push(null);
return false;
}
// Continue walking
return true;
}
async _step() {
try {
this._dir = this._dirs.pop();
this._files = await fs.promises.readdir(this._dir.path, { withFileTypes: true });
} catch (e) {
this.emit('error', e); // Uh oh...
}
}
async _walk() {
this.walking = true;
while (this._drain()) {
await this._step();
}
this.walking = false;
}
_read() {
if (!this.walking) {
this._walk();
}
}
}
stream.pipeline(new Walker('some/dir/path', 5),
fs.createWriteStream('output.txt'),
(err) => console.log('ended with', err));

Categories

Resources