How to find and resolve duplicate documents

How SparkReceipt detects duplicate receipts and documents, how to review them side by side, and how to merge documents the system didn't flag.

ST
Written by SparkReceipt TeamUpdated Aug 31, 2026

SparkReceipt checks every new document for duplicates automatically. Flagged documents get a Duplicate badge and can be reviewed in a side-by-side resolve flow where you can trash the duplicate, combine the two into one document (keeping whichever copy you choose), or mark it as not a duplicate. You can list all flagged documents with the Duplicate filter in the Documents view, and you can manually Merge any documents the system didn't flag.

How duplicates are detected

Detection runs automatically in the background shortly after a document finishes processing. A document is flagged as a possible duplicate of an earlier one in two cases:

  • Identical file — you uploaded the exact same file again (checked against your documents from the last 90 days).
  • Matching content — the documents have the same date, same amount, same currency, and a clearly similar merchant name. This catches the same receipt arriving twice in different forms, for example once scanned and once by email.

When a pair is found, the older document is treated as the original and the newer one is flagged. Nothing is deleted automatically — the new document is uploaded normally and simply marked for your review. Right after an upload you may also see a Possible duplicate dialog offering to resolve it immediately.

Duplicates that arrive by email (the same e-receipt fetched from your inbox and also forwarded manually, or the same email delivered twice) are filtered out by a separate automatic check before they ever become documents.

Finding all your duplicates

  • In the Documents browse view, set the Duplicate filter to Yes to list every document currently flagged as a duplicate. You can save this as a saved search if you review duplicates regularly.
  • Flagged documents show a Duplicate badge in lists, and an open document shows a notice: "This document may be a duplicate of …" with a Resolve duplicate button.
  • You also get a notification when a duplicate is detected.

Resolving a duplicate

Click Resolve duplicate on a flagged document (or start from the upload dialog). The resolve view shows the two documents side by side — Duplicate on one side, Original on the other — with their dates, amounts, and categories, plus a View document link for each. Each card has a Keep this one control (under Choose which document to keep); the Original is selected by default. If several duplicates are waiting, you move through them with Previous/Next ("Resolving 1 of 3 duplicates").

For each pair you have these options:

  • Trash duplicate — keeps the original and moves the flagged document to the trash. The right choice when the same receipt simply came in twice.
  • Combine — merges the two into one document: the files (images/PDFs) from the copy you're discarding move onto the copy you keep, and the discarded one is trashed. Use this when each copy has something the other lacks — for example one has the photo and the other arrived by email with the PDF. You pick which copy survives with the Keep this one control:
    • Leave the original selected (the default) and the button reads Combine with original — "Add files to original, then trash". The older document is kept and the flagged one is trashed.
    • Click Keep this one on the duplicate's card to keep the newer document instead. The button switches to Combine with duplicate — "Add files to duplicate, then trash". This is the one to use when you've already categorized, tagged, or edited the newer copy and want to keep that work.
  • Not a duplicate — keeps both documents and removes the flag. Use this for false positives, such as two genuinely separate purchases of the same amount at the same store on the same day.

Good to know about combining: only the files move onto the copy you keep. Any edits, tags, or line items on the copy you discard do not transfer — so if you already edited one of the two, keep that one as the survivor and check it afterwards. Nothing is lost permanently either way: the trashed document (and its remaining data) is in your trash.

Merging documents the system didn't flag

Detection is deliberately conservative — if the dates or amounts differ (a tip changed the total, or the scan misread a date), two documents covering the same purchase will not be flagged. You can merge them yourself:

  1. Open the Documents browse view and select the documents (at least 2).
  2. Open the batch Actions and choose Merge.
  3. In the "Choose the document to keep" dialog, pick the keeper — "Files from the other documents will be moved here. The other documents will be moved to trash."
  4. Confirm with Merge into selected.

The same Merge action is available in the mobile app's multi-select menu. As with combining, files move to the keeper while the other documents' edits and tags do not — make sure the keeper is the copy whose data you want.

This is also the way to join a two-part scan of one long receipt into a single document — see Can I scan a multi-page (2-page) invoice / Receipt? for scanning multi-page documents correctly in the first place.

Duplicates vs. bank statement matching

Matching a receipt to a bank statement transaction is a different mechanism from duplicate detection — matching links a statement line to its receipt so converting the statement doesn't double-count it. If your statements are creating documents that overlap with your receipts, the fix is matching, not duplicate resolution — see Why aren't my receipts matching bank transactions?