PDF to Excel Converter

A PDF has no table — it has text that looks like one

This is the whole difficulty, and it's worth understanding before you judge the output. A spreadsheet stores a grid: this value is in row 4, column C, full stop. A PDF stores instructions to draw glyphs at coordinates. When you look at a PDF table you are reading a grid that exists only in your head — the file itself has no rows, no columns, and no cells. Extraction means inferring the structure back from where the ink landed.

That inference works well on tables that were generated by software — bank statements, invoices, exported reports — because their columns line up to the pixel. It works less well on a table that was scanned, hand-laid-out, or built with inconsistent spacing, because the geometry it's reading is genuinely ambiguous.

The invisible characters that break column detection

Columns are found by looking for the gaps between them: a vertical corridor with no text in it is a column boundary. Simple, and it fails on a surprising number of real PDFs, because PDF renderers emit whitespace-only text items positioned inside those gaps. A space character with no glyph, sitting exactly where the corridor should be, silently bridges two columns into one.

Those items are stripped before the gaps are measured, which is why a statement whose columns kept merging elsewhere comes apart correctly here. It's a small fix for a failure that looks, from the outside, like the tool simply can't read tables.

What you'll get, and what to check

• Values arrive as text. A column of amounts comes across as amounts, but Excel may not treat them as numbers until you convert the column — currency symbols and thousands separators are part of the extracted string.
• Merged header cells are the most common casualty. A header spanning two columns has to be assigned to one of them, and the choice is a guess.
• A scanned PDF has no text at all — only a picture of text. Nothing can be extracted from it without OCR first; try OCR to get the text out, then bring it here.
• Multi-page tables come across page by page. If a table runs across pages, so will its extraction — stitch them in the spreadsheet.

When to use this and when not to

Reach for it when the data exists only as a PDF and retyping it is the alternative — a statement your bank won't export, a report from a system that only prints, a table in a paper you're citing. It will save you an hour of typing and you'll spend ten minutes checking it, which is a good trade.

Don't reach for it when the source data exists somewhere else. If whoever sent the PDF can send the .xlsx or the CSV, that file is exact and this one is inferred. Ask for the original first — it's always the better path.