Separate reported text from numeric value
Preserve the original value through extraction so a neat-looking spreadsheet does not force someone back to the PDF to recover precision.
Test an invoice with the dates, quantities or amounts you find hardest to copy accurately. ParseMill puts extracted invoice fields beside the source for review and correction before export. Open the downloaded workbook in your usual spreadsheet application and check those values there too.
Do not infer measurement accuracy or uncertainty from the number of displayed decimal places alone. A reporting convention, rounding rule or instrument display may explain them; the source must establish any stronger interpretation.
Try a precision-preservation worksheet
| Raw source | Numeric analysis value | Presentation check |
|---|---|---|
| 42.60 | 42.6 | Raw field retains final zero |
| 0.0040 | 0.004 | Raw field retains four decimal places |
| 100 | 100 | No decimal places added to raw field |
These invented values demonstrate transcription and display, not laboratory performance. A format showing two decimal places could make 100 appear as 100.00, which would add presentation detail absent from the source. Keep such formatting in a clearly labeled analysis view rather than rewriting the raw transcription.
Microsoft: Power Query data types distinguishes text from numeric types and documents numeric precision limits. Use the distinction deliberately when designing the two fields.
Check each transformation boundary
- Capture the exact reported string, including trailing zeros, sign and qualifier. Store units and source location alongside it.
- Convert a copy to a numeric type only when the notation is understood. Record any rounding rule applied for the intended analysis.
- Export and reopen the file in the actual receiving application. Compare the raw text character by character and inspect the numeric field separately.
- If a display format is required, document it as a presentation choice. Keep it from replacing the source string in future exports.
NIST: Basic uncertainty definitions describes measurement uncertainty as a property associated with a measurement result. A transcription check should not substitute decimal-place counting for the source’s uncertainty information.
Avoid repairing the wrong thing
Changing a numeric cell’s format can display trailing zeros, but it does not prove how many zeros the source contained. If the raw string was already lost, return to the source rather than reconstructing it from a convenient format. The same issue affects mixed precision: one uniform format can hide meaningful differences in reporting.
Record source corrections separately from display corrections. If the issuer changes 42.60 to 42.61, that is a different event from changing your workbook’s number format to show two decimal places.
Accept faithful transcription first
The fixture passes when raw strings remain exactly 42.60, 0.0040 and 100 after the round trip. Numerical comparisons belong in their own checks, with suitable precision and rounding rules for the task. Neither check establishes that the underlying measurements are accurate.
For the invoice trial, choose amounts and quantities with different decimal presentations. Check the source, review data and exported cells together, including any correction time. Measurement precision and uncertainty in laboratory reports need a separate document-workflow assessment.
Put it to work with ParseMill
Make one real invoice your buying test
Choose an invoice you would otherwise enter by hand. Upload it, inspect the extracted fields, correct what needs attention and download Excel. Include every checking minute in the comparison. Subscribe if the result and the time saved make sense for your recurring work.
Eligible new workspaces: 10 pages for 7 days after email verification. No card or automatic subscription.
Invoice processing: $49 USD/month · 500 PDF pages · one user. Buy only after you have checked the fit on your own invoices.
Sources and editorial notes
- Microsoft: Power Query data typesChecked 2026-10-25
- NIST: Basic uncertainty definitionsChecked 2026-10-25
Written by ParseMill. Worked examples are illustrative process examples, not customer results. Software availability can vary by subscription and region; check the linked documentation for your account. Review decisions and approvals remain with your team.