Get PDF Tables Into Excel Without Retyping

Converting a PDF table to Excel means recovering rows and columns from a page that stores no table at all — just text fragments at coordinates, sometimes with ruled lines drawn around them. Extraction tools rebuild the grid from those positions, which works well on cleanly laid-out statements and reports and fails on scans until OCR has created text to work with. This page shows the three reliable routes: Excel's built-in Get Data From PDF, the free Tabula extractor, and Acrobat's Excel export.

Published · Last updated

Why there is no one-click converter on this page

Reliable table extraction needs geometric analysis of ruling lines and text alignment plus a spreadsheet writer — engines that exist in Excel, Tabula, and Acrobat but not in anything a browser page can run trustworthily. Getting a wrong number silently into a spreadsheet is worse than no converter, so we document the tools that do it properly.

Disclosure: this page contains affiliate placeholder links. If they were live, we could earn a commission at no extra cost to you.

AFFILIATE PLACEHOLDER: Recommended desktop PDF suite (placeholder link)

How to convert a PDF table to Excel (step by step)

Method 1: Excel's Get Data From PDF (Microsoft 365 / 2021+, Windows)

  1. In Excel, choose Data → Get Data → From File → From PDF and select your PDF.
  2. The Navigator lists every detected table (Table001, Table002, …) and every whole page. Click each to preview; tick the table you want.
  3. Click Transform Data to open Power Query: promote the header row, set column types (decimal, date, text), and remove stray rows.
  4. Click Close & Load. The table arrives as a refreshable query — replace the PDF later and hit Refresh to re-import.

This is the best first stop if you have it: no extra software, and the Power Query step catches type problems before they poison your formulas.

Method 2: Tabula (free, Windows/macOS/Linux)

  1. Download Tabula from tabula.technology and open it; it runs locally in your browser at 127.0.0.1 — files are not uploaded.
  2. Import the PDF, then drag a selection rectangle around the table (or use Autodetect Tables).
  3. Choose the extraction mode: Lattice for tables with ruled cell borders, Stream for tables aligned only by whitespace.
  4. Preview, fix the selection if columns merged, and export as CSV or XLSX, then open in Excel.

Method 3: Acrobat's Excel export

  1. In Acrobat Pro choose All tools → Export a PDF → Microsoft Excel.
  2. Acrobat converts the whole document, one worksheet per page region, and OCRs scanned pages automatically — useful for mixed documents, at the cost of less control per table.

How it works

Nothing in a PDF says "this is a table." The specification has no table object for page content: what looks like a spreadsheet on the page is a set of text-showing operators placing digit runs at x,y coordinates, plus (sometimes) path operators drawing the grid lines. When Excel exported that table originally, it printed cell values at computed positions and threw away the workbook structure — formulas, cell types, number formats all gone, with formatted strings like "1,204.50" left in their place.

Extractors reverse this geometrically, in one of two ways. Lattice-style detection (Tabula's Lattice, Camelot's default) looks for the drawn ruling lines, intersects horizontal and vertical segments to reconstruct the cell grid, then assigns each text fragment to the cell whose rectangle contains it — robust even when cell text wraps, but useless on borderless tables. Stream-style detection ignores lines and instead clusters text by alignment: fragments sharing a baseline form a row, and consistent vertical gutters of whitespace across rows are read as column boundaries. Stream mode handles the borderless tables common in financial statements but is easily confused by wrapped cells and multi-line headers, which is where "two columns merged into one" errors come from.

After the grid is rebuilt, values must be re-typed in the spreadsheet sense: "1,204.50" parsed back to the number 1204.5, "03/04/2026" to a date (beware US versus European order), and identifiers like "0042" kept as text so leading zeros survive. Power Query makes this step explicit; cruder converters guess, which is why post-import checking matters. Scanned PDFs sit below all of this: an image of a table contains no text fragments to cluster, so OCR must run first — and OCR's habit of reading 5 as S or 0 as O makes verifying totals afterwards non-optional.

Worked example: a 6-page bank statement

A 6-page bank statement PDF (310 KB, one continuous transaction table, 187 rows, no cell borders) needed to become a ledger. Excel's From PDF connector detected 6 per-page tables; after appending them in Power Query, promoting headers once, and typing the amount column as decimal, the full 187-row table loaded in about 4 minutes of work — with one repair: 3 transactions with two-line descriptions had spilled their second line into a phantom row, caught because the imported debit total (14,382.20) was compared against the statement's printed total and initially disagreed by the 3 misplaced amounts. The same file in Tabula needed Stream mode (no ruled lines) and one manual column divider added between the date and description columns; export to CSV and cleanup took roughly 7 minutes. Retyping 187 rows by hand was estimated at well over an hour, with worse error odds.

Frequently asked questions

How do I convert a PDF table to Excel without retyping it?

If you have Excel from Microsoft 365 on Windows, use Data, Get Data, From File, From PDF — it detects the tables and imports them through Power Query. Otherwise the free Tabula tool lets you draw a box around the table and export CSV.

Why does Get Data From PDF not appear in my Excel?

The From PDF connector ships with Excel for Microsoft 365 and Excel 2021 or later on Windows only. On a Mac or in older versions, use Tabula, Acrobat's Excel export, or copy the table via Word's PDF import instead.

Why do all my numbers land in one Excel column?

That happens when a table is pasted or extracted as plain text lines, leaving Excel nothing to split columns on. Re-extract with a table-aware tool, or use Excel's Text to Columns with a space or tab delimiter as a rescue.

Can I get a table out of a scanned PDF into Excel?

Not directly: a scanned page is one image, and table extractors need text with coordinates. Run OCR first — Acrobat, ABBYY FineReader, or OCRmyPDF — then extract; expect to verify digits, because OCR errors in numbers are easy to miss.

Do Excel formulas come back when I convert a PDF to Excel?

No. A PDF stores only the displayed result of each cell, so a total that was =SUM(B2:B9) in the original arrives as the plain number. Rebuild key formulas after import and use them to cross-check the extracted values.

What is the best free tool for extracting PDF tables?

Tabula is the best-known free option: you draw a selection around each table and export CSV, choosing lattice mode for ruled tables or stream mode for whitespace-aligned ones. For scriptable extraction, the Python library Camelot offers the same two modes plus accuracy reports.

Related tools