Sonarr file name parser takes issues with numbers

Sonarr Version
2.0.0.4689
Mono Version
4.6.2
Synology NAS

I’ve just switched over from sickrage, and I’ve been noticing some odd behavior from the episode name parser. If any numbers exist in the episode title portion of the file name, it seems to identify them as an absolute number, and fails to import them to database when searching the disk for episodes. The parser properly identifies other episodes and adds them to the database. The trace level log shows the parser sees absolute numbers in the file name. Here are some examples of file names that failed to be parsed.

Marvel's_Agents_of_S.H.I.E.L.D_-_S03E05_-_4,722_Hours_[D376A09A].mkv

Marvel's_Daredevil_-_S02E11_-_380_[5BD619EB].mkv

85% of all episodes in this show failed to add to database
Last_Week_Tonight_with_John_Oliver_-_S03E08_-_Episode_67_[73921DBD].mkv

The_Simpsons_-_S28E17_-_22_for_30_[050E438A].mkv

I do not use absolute numbers in any of my naming conventions for north american television, even if some shows title names look like it. Filebot and sickrage had no issues with this, as they allow you to flag shows as “anime” or “using absolute numbers”. This has caused 12% of my 20TB database to be invisible to Sonarr, and I’m not sure how to proceed. Considering the fact that Sonarr itself is renaming episodes to include episode title, after the import, they appear as not on disk after the next scan.

Trace Log:

17-4-28 11:33:15.9|Debug|ImportDecisionMaker|Analyzing 1/85 files.
17-4-28 11:33:15.9|Debug|Parser|Parsing string 'Marvel's_Agents_of_S.H.I.E.L.D_-_S03E05_-_4,722_Hours_[D376A09A].mkv'
17-4-28 11:33:15.9|Trace|Parser|^(?<title>.+?)(?:[-_\W](?<![()\[!]))+(?:S?(?<season>(?<!\d+)\d{1,2}(?!\d+))(?:(?:[ex]|\W[ex]){1,2}(?<episode>(?<!\d+)\d{2}(?!\d+)))+).+?(?:[-_. ]?(?<absoluteepisode>(?<!\d+)\d{3}(?!\d+)))+.+?\[(?<subgroup>.+?)\](?:$|\.mkv)
17-4-28 11:33:15.9|Debug|Parser|Episode Parsed. Marvel's Agents of S H I E L D - S03E05 
17-4-28 11:33:15.9|Debug|Parser|Language parsed: English
17-4-28 11:33:15.9|Debug|QualityParser|Trying to parse quality for Marvel's_Agents_of_S.H.I.E.L.D_-_S03E05_-_4,722_Hours_[D376A09A].mkv
17-4-28 11:33:15.9|Debug|Parser|Quality parsed: HDTV-720p v1
17-4-28 11:33:15.9|Debug|Parser|Release Group parsed: D376A09A
17-4-28 11:33:15.9|Trace|EventAggregator|EpisodeFileDeletedEvent <~ MetadataFileService
17-4-28 11:33:15.9|Trace|EventAggregator|EpisodeFileDeletedEvent <~ OtherExtraFileService
17-4-28 11:33:15.9|Debug|ImportDecisionMaker|Size: 1743519310
17-4-28 11:33:15.9|Trace|DiskScanService|Import decisions complete for: [263365][Marvel's Agents of S.H.I.E.L.D.] [00:00:00.2821835]
17-4-28 11:33:15.9|Info|DiskScanService|Completed scanning disk for Marvel's Agents of S.H.I.E.L.D.

It’s the Hash in the file names within square brackets along with a number in the episode title (could be the series title as well in some cases) causing it.

You can use Manual Import on Wanted: Missing to tell Sonarr what the correct episode is as a workaround or not use the square brackets (obviously a big change).

Are you running a script that adds the hash after Sonarr imports it? That would cause Sonarr to have to re-parse the file and in this case not parse it properly.

Changing the parsing to take the series type into account (when known) is planned, but it’s not something we’re actively working on.

I do have a python script adding the hash with the on download trigger.

I’ll try manually importing them, there’s just a lot of files.

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.