Deduplicator - Which one will be removed

[francoisrob]

Of the deduplicator results, what is the logic behind which file will be removed and which one kept?
Is there a way to prioritise directories, metadata and file size (if perceptual hash detection was enabled),
And what would happen if I delete all where there are clusters with more than 2 files? Thanks!

[d4rken]

Of the deduplicator results, what is the logic behind which file will be removed and which one kept?

The logic is defined here:
https://github.com/d4rken-org/sdmaid-se/tree/main/app/src/main/java/eu/darken/sdmse/deduplicator/core/arbiter/checks

The general order of checks which files to keep are by descending priority:

• Prefer direct checksum matches over perception hashes
• Prefer files that are in the gallery (media provider)
• Prefer files files on primary storage
• Prefer files that are less nested
• Prefer files that are older
• Prefer files that are larger

Is there a way to prioritise directories, metadata and file size (if perceptual hash detection was enabled),

This is currently hardcoded. I want to add a way to customize this, but have not found the time to do so (and it's difficult to come up with a good UI that everyone is not overwhelmed by).

And what would happen if I delete all where there are clusters with more than 2 files? Thanks!

Based on the above criteria, one file would be picked to keep, the rest would be deleted.