What is a duplicate?
A duplicate is a file whose content matches another file’s. Not its name, not its location, not its date.
This is the single most important idea in DupliWalker, because it is the one that disagrees with how people usually look for duplicates.
Why names, dates and locations prove nothing
Consider four situations you almost certainly have on disk right now:
| What you see | What it means |
|---|---|
invoice.pdf and invoice (1).pdf | Probably the same file. Probably — the name is a hint, not evidence. |
IMG_4021.jpg and holiday-beach.jpg | Could easily be the same photo, renamed. |
| Two files with identical names in different folders | Could be two completely different documents. |
| Same name, same size, different modification date | A copy, or an edit. The date alone cannot tell you. |
A filename is something a person or a program chose. It can be changed without touching a single byte of the file, and it frequently is — by downloads, by imports, by sync tools, by “Duplicate” in a file manager. Modification dates change for reasons that have nothing to do with content, including being copied between drives.
So DupliWalker treats names and dates as hints about where to look, and content as the thing that decides.
Exact duplicates
An exact duplicate is a file whose content is identical to another’s. Two files that are exact duplicates are interchangeable: whichever you keep, you have lost nothing.
This is the case DupliWalker is built for, and it is what “duplicate” means throughout this documentation unless a page says otherwise.
Files that look alike but are not duplicates
Several kinds of near-match are not exact duplicates, and treating them as if they were is how people lose work:
- An edited copy. Same name, same origin, one paragraph different. Keeping either one is not equivalent.
- A re-encoded or re-exported file. The same photograph exported twice at the same settings can still differ byte-for-byte — different metadata, a different timestamp inside the file, a different encoder version.
- A resized or recompressed image. Visually identical at a glance, materially different as a file.
- The same document in two formats. A
.docxand its exported.pdfare not duplicates of each other in any sense DupliWalker uses.
DupliWalker can help you find candidates like these, but it will not tell you they are interchangeable, because they are not.
What this means in practice
Two habits follow from all of this:
- Do not assume that everything grouped together is safe to delete down to one copy without looking. Grouping is a finding, not a verdict. That is why comparing files is a step of its own.
- Do not dismiss a group because the names differ. A renamed copy is still a copy, and it is exactly the kind DupliWalker is best at finding and you are worst at spotting.
Next
- Dupli and Reference folders — how to tell DupliWalker which copies you intend to keep.
- Duplicate groups — how results are organized once a scan finds something.