How to find duplicate photos on an external drive without losing anything

Updated 27 October 2026.

The drive in the drawer has every photo from 2009 to 2018, some of them three times. You want to delete duplicate photos on your Mac without deleting the only copy of anything. Here is how to approach it so a mistake is always recoverable.

Why old photo drives are the hard case

Apple Photos will not help here: its Duplicates album only looks inside the Photos library, never at a folder or a drive. So you need a duplicate file finder for Mac that can scan an external volume. That is where people run into three problems.

Tools hang

Finding exact duplicates means reading file contents, not names. A careful tool compares sizes first, then hashes a small part of each candidate, then hashes the whole file. On a fast internal SSD that is quick. On a spinning USB drive, or a network share, reading hundreds of gigabytes can take hours, and a tool that fires off many parallel reads makes a spinning disk thrash and slow down further. If a scan looks frozen at 40 percent, it is often just waiting on the drive.

Tools crash

Similar-photo detection decodes every image to compare them visually. Decoding full-size images for a hundred thousand files uses a lot of memory, and some apps fall over. "Crashes just about every time" is a real complaint in recent App Store reviews of duplicate finders. Tools that decode thumbnails instead of full images, and that save their progress, cope much better.

Tools delete the wrong copy

This is the one that matters. One reviewer of a freemium duplicate finder wrote that it "goes by name, not content" and "almost deleted old family photos." On Reddit, a user summed up running a subscription finder on an external drive as a "big mistake." Whatever tool you use, the process below is designed so that this cannot cost you a photo.

Before you scan: back up

Do this even if it feels excessive. A duplicate cleanup is one of the few routine tasks that can destroy data if it goes wrong, and the drive you are cleaning is often the only copy.

1. Copy the whole drive once. Drag the drive's folders to another drive, or use Finder's duplicate, or a cloning tool. Check the copied size matches.
2. If the photos are also in Photos, make sure Time Machine has run recently and includes the Photos library.
3. Note what "original" means to you. Is it the RAW file, the largest JPEG, the one in Photos, or the oldest file date? Decide now, before a tool's default rule decides for you.

A safe order of operations

1. Scan read-only first. Run the scan and look at the results without removing anything. See how many groups there are and how big the savings would be. If it is 2 GB, you may not need to bother.
2. Check that matches are by content. Open one group and confirm the files really are the same image. If a tool pairs IMG_1023.jpg with a different IMG_1023.jpg from another camera, stop using that tool.
3. Start with exact duplicates only. They are unambiguous. Leave "similar" matching for a second pass once you trust the tool.
4. Review the keeper rule, group by group. Largest file, newest, RAW over JPEG, or the copy inside Photos. Prefer a tool that shows you which rule it applied and why, and lets you change the keeper in a group with one click.
5. Remove to the Trash, never permanently. If a tool offers "delete permanently" or "skip Trash," turn it off. The Trash costs nothing and is your undo.
6. Work in batches. Do one year, or one folder, at a time. Empty the Trash only after you have opened the folder, looked, and are sure.
7. Keep a log. Some tools write a record of what they moved and from where. If yours does not, export or screenshot the results before you confirm. It is the only way to answer "where did that file go?" a month later.

Similar photos: the second pass

Once exact duplicates are gone, most of what is left is near-duplicates: a burst of seven frames of the same moment, a photo and its resized export, a HEIC and the JPEG converted from it, the edited version and the original. These can be most of the wasted space, and they need a human eye.

• Use a tool with an adjustable similarity threshold, and start strict. Loosen it only after you have seen what the strict setting groups together.
• Look at every group side by side before you act. Two photos of a sunset a second apart are "similar" to a computer and different to you.
• Never let a tool auto-select across similar groups and remove without review. Auto-selection is fine for exact duplicates; for similar ones it is how the wrong frame goes.

Large libraries: making the scan finish

• Plug the drive in directly, not through a hub, and keep the Mac awake and on power.
• Exclude folders you know are backups of other folders. Scanning a Time Machine or clone volume alongside the original produces a duplicate for every file and tells you nothing.
• Scan subfolders separately if the whole drive is too much. Duplicates between folders can be caught in a final pass.
• Prefer tools that cache what they have already hashed, so a second scan after a crash or a sleep is fast.

After the cleanup

Keep the backup copy you made at the start until you have lived with the cleaned drive for a few weeks. Then decide whether the drive is the archive or whether the photos belong in Photos. Either way, the built-in Duplicates album is enough from then on for anything you import twice, and the comparison of Mac duplicate finders covers the tools for everything else.

How SameScan approaches this

SameScan is a Mac app in development, built around exactly this process. It scans folders, external drives, and the Photos library; matches by content; shows every group side by side with the keep-best rule and its reason; moves extras to the Trash only; and keeps a restorable log of each action. Nothing is uploaded. It is not available yet; join the launch list and we will send one email when it ships.