Duplicate Document Finder for Windows
Duplicate documents are hard to find because Office and PDF files record their own history - author, revision count, last-saved time - inside the file. Re-saving a document changes those values, and with them the checksum, so two copies of identical content no longer look identical to a duplicate file finder. Matching them means comparing document properties instead.
Duplicate File Detective finds duplicate documents both ways. Byte-identical copies are caught by content matching, exactly as they are for any other file type. Re-saved copies - the ones a checksum can never catch - are caught by comparing the properties that describe the document's content rather than its history.
Why the same document produces two different checksums
A Word document is not just its text. Alongside the content, the file stores a set of properties describing the document itself, and several of them change on every save:
- Author and last-editor - different for every person who touches the file.
- Revision number - incremented each time the document is saved.
- Total editing time - accumulates while the document is open.
- Last-saved timestamp - rewritten on every save, even when nothing changed.
A checksum covers the entire file, these values included. Open a report, change nothing, press Ctrl+S, and the file now has a different checksum from the copy sitting in the shared folder. The two documents are the same to any reader and different to any content-based comparison - which is why scanning a document library by content alone reliably under-reports the duplication in it.
The same effect shows up whenever a document changes hands. The version a colleague emailed back, the copy saved from a browser preview, the file re-exported to PDF from the same source - all carry identical content and different bytes.
How to find duplicate documents
- Start a duplicate document project. Open Duplicate File Detective, start the New Project Wizard, and choose the Duplicate Documents project type. This matches by file extension, size, and a 256-bit content checksum across common document formats - the right starting point when your copies really are copies.
- Add the folders that hold your documents. Include every location worth comparing - your documents folder, a shared drive, an archive volume, a folder of email attachments. Paths are scanned together in a single pass, so copies scattered across them are compared against each other. Network locations can be entered as UNC paths.
- Run the search and review the exact matches. Byte-identical copies are grouped first. For many document collections this alone accounts for most of the duplication, and it is the most trustworthy result the program can give you.
- Switch to document property matching for re-saved copies. For documents that were re-saved rather than copied, open Comparison Options, turn off File contents and File sizes, and enable Document properties instead.
- Choose which properties must match. Select from title, subject, page count, word count, character count, embedded creation date, and PDF document lineage. Combining several - title with page count and word count, for example - is far more reliable than matching on any one alone.
- Review every group before acting. Property matches are evidence, not proof. Use the preview pane to read the first page of each candidate, and the report columns to compare the properties the match was based on.
- Move, archive, or delete the copies you do not need. Mark the redundant copies and process them with the Duplicate File Manager - move them elsewhere, zip them for recoverability, or delete them to the Recycle Bin.
Which properties to match on
The properties fall into two groups, and the distinction is the whole idea behind this feature. Some describe the content, and survive a re-save. Others describe the file's history, and are exactly what breaks checksum matching. Only the first group is offered as matching criteria:
- Title and Subject - useful when your documents genuinely carry them; many never have them filled in.
- Page count, word count, and character count - the most broadly useful combination, because they are derived from the content and populated automatically.
- Created date (embedded) - the creation date stored inside the document, which survives copy and save-as operations that reset the file system timestamp.
- PDF document lineage - a shared document identifier linking PDFs that descend from the same source. PDF files only.
Match on several rather than one. Word count alone will pair unrelated documents that happen to be the same length; word count together with character count and page count almost never will. Note that every criterion you enable must be satisfied - the criteria narrow the results rather than widening them - and that a property missing from either file fails the comparison rather than being ignored.
What property matching will not tell you
Content matching and property matching answer different questions. A checksum match means these files are identical. A property match means these documents describe themselves identically - which is strong evidence, not proof.
- Documents from a shared template can collide. Two invoices, two forms, or two report skeletons built from the same template may match on title and page count while containing entirely different information. Always open a preview before deleting.
- Documents without properties cannot be matched this way. Plain text, scanned images saved as PDF without a text layer, and files whose properties were stripped have nothing to compare. They fail the comparison rather than being grouped.
- PDF lineage means shared ancestry, not identical content. Two revisions of the same contract share a document ID and may differ substantially. Use the PDF Revision column to see which is which.
- An edited copy is a different document. If someone revised the text, word and character counts will differ and the copies will not group - which is the correct outcome, but it means property matching finds re-saved copies rather than similar documents.
Reviewing what you found
Because property matches need confirming, the review step matters more here than with content matching. The preview pane renders the first page of a PDF and the contents of text and source files directly in the results report, so the usual question - are these actually the same document? - can normally be answered without opening anything. The report columns show the property values the match was based on, so you can see exactly which criteria tied a group together.
When you are ready to act, SmartMark can mark by age, path, or location so the copy in your archive folder is kept and the strays elsewhere are marked. There is a SmartMark option to leave at least one unmarked file in every duplicate group, and Duplicate File Detective warns you before processing if any group has all of its entries marked.
Duplicate documents on shared and server storage
Document duplication is usually worst where several people work on the same material. The same attachment gets saved into six departmental folders; a report is archived by both its author and the team that commissioned it. Duplicate File Detective scans multiple local and network paths - including UNC shares and NAS devices - in a single pass, which is what makes duplication across those locations visible at all.
On shared storage the useful output is often a report rather than a deletion. The duplicate file owners report shows who owns the duplicate documents that were found, results export to Excel, PDF, CSV, HTML, and XML, and the built-in scheduler can run the whole thing on a recurring basis and email the result - so the people who own the files can make their own decisions about them.
Duplicate Document Finder FAQ
Why do identical documents have different checksums?
Because Office and PDF files store information about themselves inside the file. Author, last-editor, revision number, total editing time, and the last-saved timestamp are all embedded in the document, and several of them change every time the file is opened and saved - even if you change nothing in the text. A checksum covers the whole file, metadata included, so two copies of identical content produce completely different checksums. This is why a normal content-based duplicate scan reports fewer duplicate documents than you know you have.
How does document property matching work?
It compares the properties that describe what is in the document rather than what has happened to it. Title, subject, page count, word count, character count, and the embedded creation date all survive a re-save, whereas the checksum does not. You choose which properties must match, and files are grouped only when every selected property is identical. Properties are read through the Windows Property System - the handlers for Office formats ship with Windows, and PDF handlers arrive with Windows 10 or later or with an installed PDF reader.
What is PDF document lineage?
PDF files can carry a shared document identifier that is preserved when the file is revised or re-saved, so copies descending from the same source document can be matched even when their content has since diverged. That last part matters: lineage tells you two PDFs share an ancestor, not that they are the same document today. The PDF Revision report column shows which revision each file represents. Treat lineage as a way of gathering related files for review rather than as a deletion-grade match.
Which document formats are supported?
Word, Excel, and PowerPoint files in both the modern XML formats (.docx, .xlsx, .pptx) and the older binary formats (.doc, .xls, .ppt), plus PDF. Office property handlers are part of Windows, so nothing extra is needed for those. PDF property support comes from Windows 10 or later, or from an installed PDF reader. A document whose properties cannot be read simply fails the comparison rather than being matched incorrectly.
Is document property matching as reliable as a checksum?
No, and it is not meant to be. A checksum match means the files are identical; a property match means the documents describe themselves the same way. Two genuinely different documents built from the same template can share a title and page count, so property matches must be reviewed before anything is deleted. Select several properties rather than one, keep the preview pane open, and treat the results as strong candidates rather than confirmed duplicates.
Can it find duplicate documents on a file server or shared drive?
Yes. Multiple local and network paths - including UNC shares and NAS devices - are scanned in a single pass, which is what makes cross-department duplication visible. The duplicate file owners report additionally shows who owns the duplicate documents it finds, and results can be exported to Excel, PDF, CSV, HTML, or XML so the cleanup can be delegated rather than performed on someone else's files.