First of all I hope this hasn’t been requested before. The search did not result in any similar request.
This is not a very pressing issue, since it can be circumvented by manual search. I had downloaded a double episode file for the season finale of Once Upon a Time. This happened to be a two-hour finale, released as separate episodes. The HDTV version was downloaded fine, no problems there. However, I had set up a profile to download WEB-DL quality once it becomes available. Most (if not all) WEB-DLs tend to be released as one file per episode. Unfortunately Sonarr doesn’t take the WEB-DLs into account for both episodes when deciding whether or not to download both files to replace the single file. It only looks at the files individually and then determines there are less episodes in the WEB-DL than in the already downloaded HDTV, and then decides against downloading the WEB-DL.
I am aware that this could be a difficult task to perform, and since it is no game breaking issue it most likely isn’t going to be added anytime soon. However if you’re already looking into similar issues it may be a good idea to work on this at the same time.
If you have any questions you need me to answer about the specifics, I’d be happy to help.
This case has a couple scenarios that need to be taken into consideration:
Grabbing of the episodes (sending to your download client)
Importing of the downloaded episode
In both cases Sonarr will only act on a release/file if it is an upgrade for all episodes in the file, whether the file contains one episode or 10, this ensures a valid file isn’t overwritten with less episodes that are available for import, causing one episode to be lost. Since Sonarr only operates on one release/file at a time it has no knowledge that another release or file could be combined with it to achieve the same state.
Its possible that we could allow grabbing of single episode releases to replace a multi-episode file, but I don’t see the import ever being automatic, its too prone to failure and would be pure guessing, which we wouldn’t want to do. I think the likelihood that we will even do the grabbing is low and even if we (the dev team) agree its something we want to do, I think it will a low priority for us.
I agree it would probably accidentally delete files you wouldn’t want to. The “issue” was mainly with grabbing. Thanks for the reply, I’ll keep an eye on GitHub.