To convert a PDF to Word with tables intact, start from the original file if you still have it: the Word document or spreadsheet that made the PDF already holds real tables. If the PDF is all you have, convert it to .docx and check four things in order: tables without rules, merged cells, right-aligned numbers and fonts. The words in a scanned PDF are read from the page first, and its tables come back as lines of text.
A PDF does not contain a table the way a Word file does. It contains ink in positions: words at points on the page, and lines drawn between some of them. A converter that hands you back a Word table has worked out where the rows and columns must be, and the quality of that guess is the whole story of this conversion.
How do you convert a PDF to Word and keep its tables?
On Whatafile it takes four steps:
- Open PDF to Word and select or drop your PDF.
- Keep DOCX under Word version. DOC is for software from before 2007, and long PDFs convert more reliably to .docx.
- Press Convert to DOCX and download the file.
- Open it in Word and compare each table with the PDF, using the four checks below.
When the PDF has real text, each table comes back as a Word table with its cells. The conversion is graded Very good, not exact, because the document is rebuilt, not traced. The free plan allows 3 conversions a day, one file at a time, each up to 25 MB, and files are deleted after 24 hours.
What breaks when a PDF table becomes a Word table?
In October 2026 we converted test PDFs holding five kinds of table, plus a scan, to Word on Whatafile and compared every cell with the original.
| Table in the PDF | What came back in Word |
|---|---|
| Rules round every cell | A Word table with all 12 cells and its borders |
| The same table with no rules | A Word table with all 12 cells and no borders |
| A heading merged across three columns and a cell merged down two rows | Both merges kept |
| No rules, with descriptions wrapping onto a second line | Six rows instead of four: each line of text became a row |
| Rules between rows only | Two column headings joined into one cell |
| A scan of the first test PDF | No table: one line of text for each row |
Each failure has one cause: the converter infers the table from positions on the page. That gives four things to check, in this order.
Tables without rules
Rules give a converter edges to find; without them it has only the gaps between words. A plain borderless table can still come through, as ours did, but wrapped text and partial rules led to extra rows and joined cells. Mend them with Merge Cells and Split Cells on the Table Layout tab (Microsoft's steps for merging and splitting cells).
Copying a table out of a PDF viewer loses even more. Copied with the PDF engine built into macOS, each row of our test table arrived as one line with single spaces between the cells, so the heading Revenue (£) looked like two cells. Word's Convert Text to Table can rebuild it, but it starts a new column at every separator you choose, such as a tab or a comma, so the cells need one between them first (Microsoft's guide to converting text to a table).
Merged cells
Simple merges came through our test. The risk runs the other way: to cover the extra rows and joined headings, the converter added merges the original never had. Compare each merged cell with the PDF, and use Unmerge Cells where two cells should be separate.
Right-aligned numbers
Alignment is positional in a PDF, so the converter infers it. Most of our right-aligned figures stayed that way, but the widest entry in a column fills its cell and gives no clue: 12,950 came back centred and 194,250.75 left-aligned. In the borderless copy, the whole money column turned left-aligned. Select the column in Word and align it right again.
Fonts you do not have
The Word file asks for the fonts the PDF used, by name: Liberation Sans and Liberation Serif in our test. If one is not on your computer, Word shows another in its place, and a wider one can push a table past the edge of the page. Install the font or switch to one you have, then choose AutoFit Contents on the Table Layout tab (Microsoft's guide to resizing a table).
Fix it before you convert, not after
If you have any access to the file that made the PDF (the original Word document, the spreadsheet, the export step in whatever produced it), go back one step and export from there instead. If that original is an RTF file, convert RTF to Word rather than the PDF: it is one step, graded Exact, and its tables arrive as Word tables without being guessed from positions on a page. Ten seconds of looking beats an hour of rebuilding a table by hand.
If the PDF is all you have, work through the four checks in order: rules and merged cells change the shape of the table, so settle those first. And if you make PDFs yourself, keep the Word file beside each one. Word to PDF is the easy direction, because the table only has to be drawn, not worked out: you can convert Word to PDF again whenever the document changes.
What about a scanned PDF?
A scanned PDF holds pictures of pages and no text, so our converter reads the English words on each page and rebuilds them as editable paragraphs, without pictures or table grids. Our scan of the three-table PDF came back as one line of text for each row. Every figure was right, but some ruled lines were read as stray characters such as | and }, so remove those before you use Convert Text to Table.
The words are read this way only when the PDF has no text at all: in our test, a scanned page inside a typed PDF came back as a picture. For a photo of a page rather than a PDF, convert JPG to Word instead: the words are read as text, in English, Urdu, Arabic, Hindi and more, and the photo is not kept.
When to send the PDF instead of Word
If the layout matters more than the editing (a form, an invoice, a certificate, anything with a signature block), the honest answer is that you do not want Word at all. You want the PDF, and you want whoever asked for a Word file to tell you which part they actually need to change.
Most requests for “the Word version” are really requests to change one sentence.
If they need a file to edit but use LibreOffice, PDF to ODT makes an OpenDocument file; our test tables and merged cells came through as they did in Word. If only the words matter, PDF to Text is graded Words only: you keep every word and lose the table, each cell on a line of its own.
If the table is headed for a web page, fix it in Word first, then convert Word to HTML: our test tables arrived as HTML tables, merged cells included.
PDF to Word questions
How do I convert a PDF to Word without losing formatting?
Export again from the original file if you have it. Otherwise convert the PDF to .docx and check its tables, merged cells, alignment and fonts against the PDF, since the document is rebuilt, not traced.
How do I copy a table from a PDF into Word?
Copying from a PDF viewer brings the words without the grid, so Word needs Convert Text to Table and a tab between the cells. When the PDF has real text, converting the whole PDF is quicker: the table arrives as a Word table.
Will my tables still be tables in Word?
Yes, when the PDF has real text: each table comes back as a Word table you can edit cell by cell. In a scanned PDF the words are recognised, but tables come back as lines of plain text.
Can I turn a scanned PDF into an editable Word document?
Yes. When a PDF has no text in it, the English words you can see on its pages are read and rebuilt as editable paragraphs, without table grids or pictures. Check names and figures afterwards, since a character can be misread.
Should I choose .doc or .docx?
Choose .docx unless you need .doc for Word 2003 or a system that only accepts .doc. Word has used .docx since 2007, and long documents convert more reliably to it.
Is there a page or file size limit?
A PDF can be up to 2,000 pages. On the free plan each file can be up to 25 MB, with 3 conversions a day; on Pro, files can be up to 100 MB.