logoDenoby example



Piping Streams


Deno implements web-standard streams which comes with many advantages. One of the most useful features of streams is how they can be piped to and from different sources.

A common usecase for streams is downloading files without buffering the whole file in memory. First, we open a file we want to write to
const download = await Deno.open("example.html", { create: true, write: true });
Now let's make a request to a web page
const req = await fetch("https://examples.deno.land");
We can pipe this response straight into the file. In a more realistic example, we would handle a bad request, but here we'll just use the nullish coalescing operator. Instead of opening the file and manually piping the fetch body to it, we could also just use `Deno.writeFile`.
Pipes can connect a lot of things. For example, we could pipe the file we just downloaded into stdout. We could also pipe our streams through transformers to get a more interesting result. In this case, we will pipe our file through a stream which will highlight all "<" characters in the terminal.
First we will import a utility from the standard library to help us with this.
import { bgBrightYellow } from "https://deno.land/std@0.207.0/fmt/colors.ts";
Then we will create a transform stream utility class
class HighlightTransformStream extends TransformStream<string, string> {
  constructor() {
      transform: (chunk, controller) => {
        controller.enqueue(chunk.replaceAll("<", bgBrightYellow("<")));
Let's open our file for reading
const example = await Deno.open("example.html", { read: true });
Now we can pipe the result from the file, into a TextDecoderStream, into our custom transform class, back through a TextEncoderStream, and finally into stdout
await example.readable
  .pipeThrough(new TextDecoderStream())
  .pipeThrough(new HighlightTransformStream())
  .pipeThrough(new TextEncoderStream())

Run this example locally using the Deno CLI:

deno run --allow-net --allow-read --allow-write https://examples.deno.land/piping-streams.ts