Review repeated values without assuming the source match is correct

Stop searching the PDF for every repeated amount

Suppose the amount 120.00 appears on three invoice lines. Searching for 120.00 gives you three hits and another decision: which one belongs to this row? Repeat that across a monthly batch and the checking becomes part of the cost of building the spreadsheet.

ParseMill lets you select an extracted invoice field and inspect its source beside the data. Check the line description and column label, correct the field if needed, and export the reviewed result. The source highlight gives you a place to check; your review establishes whether the association is right.

The same matching problem appears in reports with repeated measurements. Use the example below to see why a matching number alone is insufficient. For your buying decision, test the invoice workflow with a real supplier invoice that contains repeated prices or quantities.

Try a repeated-value fixture

All three synthetic results look identical. Matching on result and unit cannot distinguish them. Matching on sample identifier alone also fails because S11 appears in two reports. A reliable review must retain report identity and the exact row context.

Microsoft: Document layout analysis includes row and column information with source regions. Use such coordinates to narrow the source location, then inspect the visible identifier and heading rather than trusting a search hit.

Build a source-match record

1. Record the output row key, source file version, physical page, visible identifier and field name. Keep the expected source locator with the check.
2. Open the source and inspect the chosen cell plus the identifier to its left or above. Check section-level context if identifiers repeat.
3. Compare neighboring records to detect an offset. A one-row shift can preserve every common value while assigning uncommon values incorrectly.
4. Mark ambiguous matches unresolved and describe what additional evidence would distinguish them. Do not choose the first matching text occurrence automatically.

Microsoft: Table.AddIndexColumn provides positional index values. A preserved source position can help distinguish repeated rows, provided it was assigned before sorting or combining changed their order.

Use a discriminating review sample

Include both repeated and uncommon values in a test set. Repeated values expose weak matching rules; uncommon values help reveal shifts that a run of zeros can hide. Inspect transitions where the result changes, not only a convenient middle row.

Do not deduplicate identical values during the review. Two equal results may be legitimate separate observations. If duplicate records are suspected, compare their full source identities and document the evidence for removal.

Confirm the correct association

The fixture passes only when each output row resolves to its own report and source row. Three matching zeros are insufficient evidence of correctness. Preserve the match record with the reviewed output revision so later sorting cannot silently invalidate the signoff. Check the evidence columns again whenever rows move between worksheets.

In an invoice trial, include two suppliers with repeated prices or quantities. Check whether you can find, verify and correct each relevant field quickly enough to beat your current process. Broader report extraction needs a separate fit assessment.