Node.js: Count the number of lines in a file

Question

I have large text files, which range between 30MB and 10GB. How can I count the number of lines in a file using Node.js?

I have these limitations:

• The entire file does not need to be written to memory
• A child process is not required to perform the task

Answer 1

We can use indexOf to let the VM find the newlines:

function countFileLines(filePath){   return new Promise((resolve, reject) => {   let lineCount = 0;   fs.createReadStream(filePath)     .on("data", (buffer) => {       let idx = -1;       lineCount--; // Because the loop will run once for idx=-1       do {         idx = buffer.indexOf(10, idx+1);         lineCount++;       } while (idx !== -1);     }).on("end", () => {       resolve(lineCount);     }).on("error", reject);   }); }; 

What this solution does is that it finds the position of the first newline using .indexOf. It increments lineCount, then it finds the next position. The second parameter to .indexOf tells where to start looking for newlines. This way we are jumping over large chunks of the buffer. The while loop will run once for every newline, plus one.

We are letting the Node runtime do the searching for us which is implemented on a lower level and should be faster.

On my system this is about twice as fast as running a for loop over the buffer length on a large file (111 MB).

Answer 2

solution without using wc:

var i; var count = 0; require('fs').createReadStream(process.argv[2])   .on('data', function(chunk) {     for (i=0; i < chunk.length; ++i)       if (chunk[i] == 10) count++;   })   .on('end', function() {     console.log(count);   }); 

it's slower, but not that much you might expect - 0.6s for 140M+ file including node.js loading & startup time

>time node countlines.js video.mp4  619643  real    0m0.614s user    0m0.489s sys 0m0.132s  >time wc -l video.mp4  619643 video.mp4 real    0m0.133s user    0m0.108s sys 0m0.024s  >wc -c video.mp4 144681406  video.mp4