i know that a website can be accessed through its name/url and also its unique ip/dns address. especially in a scenario of keyword blocking, a blocked website can be accessed through its ip address.
is there such a way to represent a whole web link of, for example a file, in ip/dns or some other coded form, so that file extension blocking may be beaten.

In some cases it can be accessed through its IP address but not its host name, but I don't see how that factors in to file extension blocking. As far as I know the extension based downloading does deep packet inspection on the HTTP header and looks at the file name provided by the webserver.

Here is an example of the header:

There may be another way around it but I don't think DNS/Hostnames will be the solution you are looking for.