How to Convert PDF to Excel: 5 Methods That Preserve Your Tables

By Michael Peters · Published July 28, 2026 · 7 min read

← Back to Blog

Why Convert PDF to Excel?

Data arrives as PDF far more often than it should: bank statements, supplier invoices, financial reports, published statistics, lab results, price lists. The numbers are right there on the page — but a PDF is a final-form document, and the moment you need to sort the rows, sum a column, or feed the figures into a model, you need them in a spreadsheet. Retyping is slow and introduces errors; a good conversion gets the data into Excel with its structure intact.

Common scenarios include extracting transaction tables from bank and card statements for reconciliation, pulling line items from invoices into accounting workflows, reusing published data tables from reports and papers, and recovering a spreadsheet when only the PDF version survives.

Why Tables Are the Hard Part

PDF has no concept of a table. What looks like a neat grid of rows and columns is stored as positioned text fragments and, sometimes, ruled lines — there is no metadata saying “this is a cell, this is a header, these values belong to the same column”. (Our guide to PDF file structure explains how page content is actually stored.) Every conversion tool must therefore infer the table structure from the visual layout, and that inference is where conversions succeed or fail.

Converts well: tables with clear ruled borders, one value per cell, consistent columns, and real text (selectable in a PDF viewer).

May need manual cleanup: borderless tables (the converter has to guess column boundaries from whitespace), merged or spanned cells, multi-line cell content, tables that continue across pages, and currency or percentage symbols that end up in their own columns.

Cannot be converted directly: scanned documents, where the “table” is a picture. These need OCR first — covered in Method 4 below.

Method 1 — Adobe Acrobat Pro (Export PDF)

Adobe Acrobat Pro provides the highest-fidelity conversion for digitally created PDFs. Open the PDF in Acrobat, then choose File > Export a PDF > Spreadsheet > Microsoft Excel Workbook (.xlsx). Acrobat analyses the page content, detects tabular regions, and rebuilds them as spreadsheet cells.

Two settings are worth checking before you export. First, Acrobat can create one worksheet per table or one worksheet per page — per-table is usually what you want for reports containing several distinct tables. Second, set the decimal and thousands separators to match the document’s regional format. A European PDF using 1.234,56 imported with UK settings will silently produce wrong numbers, which is far worse than an obviously broken layout.

For scanned PDFs, Acrobat automatically runs OCR before exporting if the page contains only image content. If you only need one table from a long document, there is also a quicker route: select the table with the Snapshot tool or drag-select the text, copy, and paste into Excel — Acrobat’s clipboard preserves basic table structure surprisingly well for simple grids.

Method 2 — Excel’s Built-in PDF Import (Power Query)

Since 2020, Excel for Microsoft 365 on Windows has been able to import PDF data directly — no Acrobat required. Choose Data > Get Data > From File > From PDF and select the file. The Navigator pane lists everything Excel’s Power Query engine has detected: each table (Table001, Table002…) and each full page (Page001…). Preview a table, then either Load it straight into the worksheet or click Transform Data to open the Power Query Editor and fix column types, remove header rows, or filter out totals before loading.

The killer feature is that the import is a refreshable query, not a one-off paste. If you receive the same report every month, save the new PDF over the old file and click Refresh — the data updates through the same query, transformations included. For recurring workflows this beats every other method in this article.

The limitation is availability: PDF import requires a Microsoft 365 subscription on Windows. It is not present in the perpetual Excel 2019/2021 licences or in Excel for Mac.

Method 3 — Free Online Converters

For a quick one-off conversion without installing anything, an online converter is the practical choice. Mapsoft’s PDF Hub converts PDF to .xlsx for free — the file is processed on Mapsoft’s own UK servers and deleted automatically within 24 hours. Upload the PDF, click Convert to Excel, and download the workbook.

With any online converter, weigh the privacy implications: the document leaves your machine. For bank statements, payroll data, or anything under NDA, prefer a local tool — Acrobat, Excel’s built-in import, or a desktop application such as Mapsoft’s PdfHub Local, which never uploads files.

Method 4 — Scanned PDFs: OCR First

A scanned PDF contains a photograph of a table, not text, so no converter can extract data from it until OCR (optical character recognition) has produced a text layer. Acrobat Pro handles this automatically during export; alternatively, run OCR as a separate first step — Mapsoft’s PDF Hub OCR tool is a free browser-based option, and our guide to OCR in Adobe Acrobat covers the desktop workflow in detail.

Set expectations honestly: on a clean 300 dpi scan, good OCR reaches roughly 95–97% character accuracy. That sounds high until you consider what the remaining errors look like in a spreadsheet — 0 read as O, 1 as l, 5 as S, a decimal point missed entirely. In prose these are typos; in a financial table they are wrong numbers that look plausible. Three rules for OCR’d data:

• Start from the best scan you can get. 300 dpi, straight, high-contrast. Re-scanning a document is often faster than fixing a bad extraction.
• Check totals against the document. If the PDF prints column totals, sum the extracted column in Excel and compare — it is the fastest way to catch a single corrupted digit.
• Spot-check a sample of rows against the original for any dataset that feeds decisions, regardless of which tool produced it.

Method 5 — Programmatic Extraction

For batch conversion — a folder of monthly statements, invoices arriving daily into a processing pipeline — a programmatic approach is the right tool.

Adobe PDF Services API. Adobe’s cloud API includes an Export PDF operation that produces .xlsx using the same conversion engine as Acrobat Pro. You send the PDF via a REST call and receive the workbook in response; SDKs are available for Node.js, Java, .NET, and Python. Pricing is per document with a free tier.

Python table-extraction libraries. camelot-py (which borrows its name from the Adobe research project that gave us PDF — see our history of Project Camelot) and tabula-py both specialise in extracting tables from text-based PDFs into pandas DataFrames, which export to .xlsx in one line. pdfplumber offers lower-level control when tables are irregular. Camelot’s two modes are worth knowing: lattice for tables with ruled lines, stream for borderless tables aligned by whitespace — choosing the right mode is most of the battle.

Commercial libraries. Toolkits such as Aspose.PDF and Syncfusion provide PDF-to-Excel conversion for .NET and Java applications running entirely on your own infrastructure — relevant where data residency rules out cloud APIs.

None of these handle scanned documents by themselves; pair them with an OCR step (Tesseract, or a cloud OCR service) for image-based input.

After the Conversion: Cleaning Up the Data

Whichever method you use, budget a few minutes for cleanup before trusting the workbook:

• Numbers stored as text. The most common post-conversion problem. Symptoms: left-aligned figures, green corner warnings, and SUM() returning 0. Fix with Excel’s convert-to-number prompt, or VALUE(), or Power Query’s column type settings.
• Currency symbols and units. £1,234.56 may arrive as text, or the symbol may land in its own column. Strip symbols and re-apply them as number formatting instead.
• Dates and regional formats. Check that 04/07/2026 was interpreted with the day/month order the document intended, and that decimal commas were not read as thousands separators.
• Merged and multi-line cells. A cell whose content wrapped over two lines in the PDF sometimes becomes two rows. If a five-row table arrives as twelve rows, this is why.
• No formulas. A PDF stores only displayed values, so calculated columns arrive as static numbers. Re-create the formulas if the sheet needs to stay live.

Tool Comparison: Which Method When

The honest production advice mirrors what we said about PDF to Word: if the table is the document — financial statements, regulatory data, anything where a shifted digit matters — try to obtain the source spreadsheet or CSV before resorting to conversion. Extraction from PDF is the right tool when the source is genuinely unavailable, and with the checks above it can be done safely.