Detect a repeated page in a combined PDF

Verify duplication in the document itself

Check the source sequence so the final workbook reflects the records you received rather than the number of pages in the bundle.

For multipage invoices, use ParseMill to review extracted invoice data beside the source and export separate summary and line-item sheets. Include a page break in your trial and check that every required line made it through. Other report types need their own fit assessment.

Keep the received PDF unchanged. Work on an authorized copy and record both physical page positions, their printed labels and distinctive content. A printed page number repeated across two report sections is not sufficient evidence of duplicate content. Look for the report identity, row sequence, headings, footnotes and revision marks together.

Use a page-identity review sheet

Adobe: Renumber pages in PDFs distinguishes PDF page labeling from the page sequence. Microsoft: Table.Distinct describes row duplication checks, which should not substitute for inspecting document identity.

1. List the suspect physical positions and their printed labels. Add the report identifier and first and last visible record on each page.
2. Compare the full visible pages, including notes and small revision marks. Record any difference, even when the main numeric table matches.
3. Decide whether the second page is an accidental repeat, a revised page or a separate report using reused numbering. Ask the source owner when the relationship is unclear.
4. If exclusion is authorized, retain the original file and mark which physical page contribution was omitted from the working register. Keep its counterpart reference and decision reason.

Distinguish repeated content from a revised page

Synthetic example: this page inventory demonstrates three superficially similar situations.

Positions 4 and 5 may represent one page accidentally inserted twice. Position 6 must not be treated as another identical copy merely because the record identifiers match. The changed revision and value require a version decision. A row-level deduplication based only on A21–A30 could discard the corrected observation or mix revisions into a register that never existed in the source.

Do not let a hash answer a different question

A file hash identifies a complete byte sequence, not whether two physical PDF pages mean the same thing. Page images or extracted text can help identify candidates, but differences in rendering, hidden text, annotations or compression can complicate comparisons. A matching text block is a prompt to inspect, not a complete document assembly decision.

Avoid deleting repeated observations simply because an amount, sample identifier or description appears twice. Replicates and repeated charges can be real. Where the source intentionally repeats a summary page for different recipients or sections, document whether your register’s scope calls for one representation or two contextual records.

Reconcile the exclusion and preserve the received evidence

After an authorized exclusion, compare the retained row identities against the chosen page contribution. Record the original page count, excluded physical positions, retained page count and any changed working-copy numbering. Confirm that all source links still refer to the received PDF rather than silently shifting after a page removal.

If the duplicate cannot be established, keep both contributions marked for review. The multipage guide provides the surrounding coverage checks. Do not present this page-identity exercise as an automatic ParseMill duplicate-page feature; it is a manual review method for protecting a combined register.