Query parameters not compared independent of order

[linuxdaemon]

In a majority of cases URL('http://example.com/?foo=bar&test=1') should be equivalent to URL('http://example.com/?test=1&foo=bar'), but they are treated as unique and there does not seem to be any extra methods for an order-independent comparison. I've had to do something similar to url.with_query(MultiDict(sorted(url.query.items()))) to both URL objects before comparing them.

A method on the URL object to do this sort of comparison would be quite useful, or maybe even implementing it in the normal comparison methods.

[aio-libs-bot]

GitMate.io thinks possibly related issues are #245 (Query parameters didn't get fully encoded), #34 (Incorrect parsing of query parameters with %3B (;) inside), #73 (Unexpected compare behaviour), #33 (Parse URL parameter), and #102 (Document encoded parameter).

[asvetlov]

By the standard, order or query values does matter.
Moreover, URL('http://example.com/?foo=bar&foo=baz') and URL('http://example.com/?foo=baz&foo=bar') are different URLs.
The default behavior should be not changed.

If sorting by order works for you -- that's fine. You can do it manually without the library modification.
If you want to propose a new comparison method -- please elaborate the new API first. How should it look like?

[linuxdaemon]

I was thinking a method like URL.unify() to 'unify' the URL format by converting it to a standard form, with like empty or default url components either always added or always removed, ensuring a consistent order with the parameters, stuff like that. My main use case for this so far is in testing a function which generates a URL with query parameters formed from a dict, as testing on different python versions produces different orders.

[asvetlov]

I'm sorry, but URL.unify() is out of the project scope.
You can write a helper for dict sorting in your project easy though.