How to get a table out of a PDF and into Excel or Word

Why tables are the hardest part of a PDF

A table that looks neat on screen is often not a table inside the file. Microsoft explains that most PDFs don't store structure such as paragraphs, tables or columns, and that a PDF "stores a table as a set of lines without any relationship to the content inside the table cells." Any tool that pulls a table out has to guess where each cell starts and ends.

That is why copy and paste so often gives you one long column of numbers, or every row squashed into a single cell. Picking the right route for your file saves most of that cleanup.

Step 1: Check whether the PDF has real text

Open the PDF and try to select one number in the table with your cursor.

• You can select individual words and numbers: the PDF has a text layer. Use Route A.

• Nothing highlights, or the whole page highlights as one block: the page is an image, usually a scan or a photo. Use Route B, or run text recognition on the PDF first so it has a text layer.

Route A: Excel's PDF import (for PDFs with real text)

Excel can read tables straight out of a PDF using Power Query, its data import tool. Microsoft lists this for Excel for Microsoft 365 and recent desktop versions. The steps from Microsoft's help page:

1. Select Data → Get Data → From File → From PDF.

2. Choose your PDF and select Open.

3. The Navigator window lists the tables Excel found. Click each one to preview it.

4. Select Load to drop the table into a sheet, or Transform Data to clean it in Power Query first.

A few things that help:

• Pick a table, not a page, when you can. The Navigator may offer both. A detected table is usually cleaner than a whole page.

• Tables that run across pages. Microsoft's connector documentation says similar tables on consecutive pages are combined into one table by default. Check that the join didn't merge two different tables.

• Clean headers in Power Query. If the first row of data appears as a header, or the header row is split across two rows, fix it in the editor before loading.

• Missing option? Microsoft notes the PDF connector needs .NET Framework 4.5 or later on Windows. If you don't see From PDF, check your Excel version.

Route B: Excel's Data from Picture (for scans, photos and screenshots)

If the table is an image, Excel for Microsoft 365 has a feature called Data from Picture. Microsoft's steps:

1. For an image file, select Data → From Picture → Picture From File.

2. For a screenshot, capture just the table (on Windows, Windows logo key + Shift + S), then select Data → From Picture → Picture From Clipboard.

3. Excel analyzes the image. Review the results, correct anything it got wrong, then select Insert Data.

Microsoft's own tips matter here: the image should show only the data you want, so crop it, and it should be taken head-on rather than at an angle. Microsoft also lists which character sets are supported. English, French, German and Spanish are on the list, along with a number of other mostly European languages. If your table is in another script, expect weaker results.

On Windows, Microsoft says Data from Picture needs Windows 10 version 1903 or later, or Windows 11.

Route C: Into Word instead of Excel

Sometimes you want the table in a Word document, not a spreadsheet.

• Open the PDF in Word. Word makes a converted copy and leaves your original PDF untouched. Microsoft says this works best on files that are mostly text, and lists "tables with cell spacing" among the elements that don't convert well.

• Rebuild a broken table. If the table arrives as lines of text, you can turn it back into a table. Microsoft's steps: make sure columns are separated by a consistent character (tabs work if the text contains commas), select the text, then choose Insert → Table → Convert Text to Table and set the number of columns.

• Go through Excel. For number-heavy tables, importing into Excel first and then pasting into Word often gives a cleaner grid than converting the whole PDF.

If the rest of the document also falls apart in Word, our guide on why PDF to Word conversions mess up formatting covers columns, headers and footnotes.

Check the numbers before you use them

This is the step people skip, and it is where most damage happens. A table can look right and still be wrong.

• Add up a column. If the PDF shows a total, sum the imported column and compare. A mismatch points to a missing or misread row.

• Count the rows. Compare the number of rows in Excel with the PDF, page by page.

• Look for shifted cells. One empty cell in the original can push every value in that row one column to the left.

• Check numbers stored as text. If Excel left-aligns numbers or won't sum them, they may have come in as text. Look for stray spaces, currency symbols or footnote markers stuck to the number.

• Watch the look-alikes. On scans, 0 and O, 1 and l, 5 and S, and 8 and B are common mix-ups. A decimal point read as a comma can turn 1.50 into 150.

• Check dates. A date like 03/04/2026 can mean March 4 or April 3. Confirm which format the document uses before Excel converts it.

A quick sample

Here is a synthetic example. A supplier sends "Northgate Supplies price list (sample)" as a two-page PDF with a five-column table: item, code, quantity, unit price, total.

1. The text selects, so you use From PDF. The Navigator shows Table001 and Table002.

2. You load Table001 and see the header row repeated halfway down, where the table crossed the page.

3. You delete the repeated header in Power Query, then sum the total column. It is off by one line, so you find a row that wrapped onto two lines in the PDF and was split in two.

Ten minutes of checking beats sending a quote built on a wrong total.

When the free routes aren't enough

Excel's import tools work well on clean, digital PDFs and simple images. They struggle with borderless tables, merged cells, tables inside multi-column pages, and scans in mixed scripts.

Need the whole document, not just one table?

EasyToDecode is launching soon with a converter for scanned PDFs and photos. It is built to rebuild tables as real Word tables, keep columns and headings in place, and flag pages that don't closely match the original so you know where to look. It will accept PDFs and JPEG, PNG or WebP images up to 10 MB. See how it works or join the waitlist to hear when it opens.