Facebook Crawler infers different og:url than the one specified in the metatag

Question

I am trying to collect the engagement data for this page: http://www.nature.com/news/the-future-of-dna-sequencing-1.22787

While working on my code I realized that Facebook returns different Open Graph objects for different variations of the URL which should not happen as the page provides a canonical URL.

Example:

1. http://www.nature.com/news/the-future-of-dna-sequencing-1.22787
2. http://www.nature.com/news/the-future-of-dna-sequencing-1.22787/
3. https://www.nature.com/news/the-future-of-dna-sequencing-1.22787
4. https://www.nature.com/news/the-future-of-dna-sequencing-1.22787/

resolve to different Open Graph objects.

Upon inspection in the Open Graph Object Debugger I saw that the inferred canonical URL had additional parameters, e.g., http://www.nature.com/news/the-future-of-dna-sequencing-1.22787?error=cookies_not_supported&code=4b78fd75-8cad-4456-a877-be0c9c1b9108

---

While I am aware that the redirects failed because of some cookie action going on, the last part of the new URL code=4b78fd75-8cad-4456-a877-be0c9c1b9108 varies for each of the 4 URL variations, which then leads to different OG objects.

Can somebody help me out here and explain why Facebook is inferring a new canonical URL (with added parameters) even though the original link had provided one?