Regular Expression for Route Params in Express.js

Question

As per various docs and blogs, as well as this question and others, I'm aware that one can validate route parameters by using regular expressions. This, however, has had me searching for roughly an hour and a half:

app.get('/api/:url(/^http:\/\/(.+?)\.(.+?)$/)', (req, res) => {
  // do stuff with req.params.url
});

Where every time I run the server locally and enter localhost://8000/api/http://www.google.com, the response is Cannot GET /new/http://www.google.com

I know the Regular Expression does what I want it to do, because :

/^http:\/\/(.+?)\.(.+?)$/.test('http://www.google.com');

... returns true.

I've also tried changing the route string to look like...

'/api/:url(^http:\/\/(.+?)\.(.+?)$)'

'/api/:url(http:\/\/(.+?)\.(.+?))'

'/api/:url/^http:\/\/(.+?)\.(.+?)$/'

And several other formats that look anything remotely like the examples given in Stack Overflow threads and blogs (the Express docs have absolutely no examples of a Regular Expression used outside of the first route param). Perhaps there are some formatting constraints that I'm unaware of, or a limitation to RegExp support in Express? It'd be great if the docs said anything at all.

I appreciate any help I can get here.

Is it a typo or do you have `localhost` defined as a protocol? Shouldn't it be `http://localhost:8000/api/...` ? — Mihai Potra, Jul 02 '16 at 08:49
That's just what I'm putting in the address bar of the browser, it really doesn't matter whether I include the protocol or not. Regardless that's not what I was asking about. — Christopher Ronning, Jul 02 '16 at 13:06

score 15 · Accepted Answer · answered Jul 02 '16 at 18:22

Your regular expression is wrong, for what you're trying to achieve.

TL;DR Use this one:

app.get('/api/:url(https?:\/\/?[\da-z\.-]+\.[a-z\.]{2,6}\/?)', (req, res) => {
});

Some explaining:

Firstly, path-to-regexp, the module used by express does exactly what its name says: converts a path to a regular expression. As such, the ^ and $ anchors are either illegal or can be interpreted differently when placed inside your :url() contents (path-to-regexp already uses them). Also, do not include the slashes used in JavaScript to identify RegExp objects, but only the expression content.

Here's how express sees your regular expressions:

path: '/api/:url(/^http://(.+?).(.+?)$/)'
regexp: /^\/api\/(?:(\/^http:\/\/(?:\.+?))\.(\.+?)$\/)\/?$/i

path: '/api/:url(^http://(.+?).(.+?)$)'
regexp: /^\/api\/(?:(^http:\/\/(?:\.+?))\.(\.+?)$)\/?$/i

path: '/api/:url(http://(.+?).(.+?))'
regexp: /^\/api\/(?:(http:\/\/(?:\.+?))\.(\.+?))\/?$/i

path: '/api/:url/^http://(.+?).(.+?)$/
regexp: /^\/api\/(?:([^\/]+?))\/^http:\/\/(?:\.+?)\.(\.+?)$\/?$/i

And, using a code snippet like this, you can see that none match:

const url = 'http://www.google.com';
for (let layer of app._router.stack) {
  if (layer.name === 'bound dispatch') {
    console.log(layer.regexp + ': matches', url, '=', layer.regexp.test(url));
  }
}

It's also advisable to avoid using capturing parentheses in an already named capture.

Example (notice the extra parentheses around the scheme of the URI, and how they change the req.params.url value):

app.get('/api/:url((https?:\/\/)?[\da-z\.-]+\.[a-z\.]{2,6}\/?)', (req, res) => {})

> req.params: { '0': 'http://', url: 'http://' }

Later edit: Please note, this entire post is only about string routes that contain regular expression within, not routes defined with RegExp objects.

score 0 · Answer 2 · answered Jul 02 '16 at 12:02

0

Try double backslash

app.get('/api/:url(/^http:\\/\\/(.+?)\\.(.+?)$/)', ...

answered Jul 02 '16 at 12:02

Aikon Mogwai

4,954
2
18
31

All that does is escape the backslash characters. Those backslash characters were put there to escape the forward-slash characters, to match a correct url. – Christopher Ronning Jul 02 '16 at 13:05
humm. That works. Any clue why that one works while the other accurate Regex does not work in this context? – Christopher Ronning Jul 02 '16 at 14:01
I should clarify: I want the one I wrote because it is more specific, and doesn't allow for things like extra slashes. – Christopher Ronning Jul 02 '16 at 14:06
The result of Express-code glance is [path-to-regexp](https://github.com/component/path-to-regexp) (express used this library to routing) – Aikon Mogwai Jul 08 '16 at 16:36

score 0 · Answer 3 · answered Jul 02 '16 at 14:00

0

I don't think you can get it to work with named parameters, but you can pass a regular expression directly:

app.get(/^\/api\/http:\/\/(.+?)\.(.+?)$/, ...);

This will yield (as req.params):

{ '0': 'www', '1': 'google.com' }

answered Jul 02 '16 at 14:00

robertklep

198,204
35
394
381

Regular Expression for Route Params in Express.js

3 Answers3