PDF to Excel: tables out, guesswork flagged
Extract tables from text-based PDFs into a spreadsheet. Ruled tables with consistent columns do well; merged headers, wrapped cells and scans do not — and the preview shows what was detected before you export.
Why table extraction is a guess
In a spreadsheet, a cell is a real object with boundaries. In a PDF there is no such thing: a table is a set of horizontal and vertical rules plus text placed at coordinates that happen to line up. The converter has to decide where one cell ends and the next begins, usually by looking for ruling lines first and column gaps second.
That works well for a bank statement or a price list. It works badly for a financial table with a three-row merged header, and not at all for a photograph of a table. The honest fix is a preview that shows the detected grid, so you can see the guess before you build a pivot table on top of it.
What converts well
- Good: statements, invoices and price lists with visible ruling lines.
- Good: tables with a single header row and consistent column widths.
- Poor: merged cells, sub-headers and cells containing several lines of text.
- Poor: tables that continue across page breaks — each page becomes its own block.
- Impossible: scanned tables; there is no text to extract.
How to convert
- Add the PDFThe preview marks the regions detected as tables, page by page.
- Choose tables only or all textTables only keeps the grid; all text also captures paragraphs outside tables, which is useful for statements with notes.
- Export and verifyOpen the .xlsx, spot-check totals against the PDF, and expect to merge header cells by hand.
| Item | Behaviour |
|---|---|
| Input | Text-based PDF with ruled or aligned tables |
| Output | .xlsx, one sheet per page by default |
| Cell values | Text and numbers as displayed — no formulas, no live links |
| Merged cells | Flattened; repeated headers are often duplicated per row |
| Scanned tables | Not supported — requires OCR plus table detection |
FAQ
Are formulas preserved?
No. A PDF only stores the calculated result, never the formula behind it. You get values in cells; rebuild any calculations yourself.
Can it extract several tables from one page?
Yes, when they are visually separate and ruled. They usually land in the same sheet, separated by blank rows, so you can split them after import.
Why are numbers arriving as text?
Because the PDF has no number type — “1,234.00” is a string of characters. Excel may or may not coerce it. If it does not, use Data → Text to Columns or a quick VALUE() formula.