How to keep tables when converting PDF to Word
Last reviewed: 2026-10-25
Tables are where PDF to Word conversion goes wrong most often. The reason is simple once you know it, and it tells you in advance which tables will survive.
Why tables are hard
A PDF does not know it contains a table. It stores pieces of text at positions on the page and, separately, lines drawn on the page. A converter has to guess that a grid of lines with text inside it is a table. The pdf2docx engine we use for normal PDFs works exactly this way: its documentation describes parsing “horizontal and vertical lines surrounding text” as a table.
So the single best predictor is: can you see lines between the cells?
What we measured
Our test invoice had three table-like areas. Here is what happened to each (full details on the test results page):
| In the PDF | In Word |
|---|
| Line items, with lines around every cell (9 rows, 6 columns) | A real Word table, all 9 rows and 6 columns, every number in the right cell. All columns came out the same width. |
| From / Bill to / Date block, no lines | A small 3-column table with some cells centred instead of left-aligned. |
| Delivery schedule, no lines (header row + 3 rows) | The 3 data rows became a table without borders; the bold header row was left outside it as one line, so headings no longer lined up with columns. |
| The same invoice, scanned | No Word tables. Each row came back as a line of text with tabs between cells, occasionally with a “|” where a cell border was. |
The Greek form’s bordered table also stayed a proper Word table, with all four rows intact.
What the converter does for tables
- Tables that would be wider than the Word page are scaled down to fit inside the margins. This was added after a real document had 6 of 24 tables running off the page, one by about 6.8 inches.
- Row heights are set as a minimum rather than a fixed height, so a row can grow when its text wraps instead of overlapping the next row.
- Word is given an explicit table width, because without it Word was found to ignore the narrowed column widths.
Convert a PDF now →
Free, no sign-up, files deleted as soon as you download the result
Repairing a table in Word
- Columns the wrong width: click in the table, then drag the column borders, or use Layout → AutoFit → AutoFit Contents to let Word size columns to their text.
- Header row outside the table: put the cursor in the first table row and insert a row above, then copy the header words into it. For a long table, turn on “Repeat Header Rows” so the header shows on every page.
- Rows that came out as tab-separated text (common with scanned tables): select those lines and use Insert → Table → Convert Text to Table, with tabs as the separator. Check the column count Word suggests before you confirm; a stray “|” or a missing tab will shift a row.
- Need the numbers in a spreadsheet? Copy the Word table into Excel or Google Sheets. Check that amounts arrive as numbers, not text, before adding them up.
Before you convert
- If you made the PDF yourself, export it again with table borders switched on. It costs nothing and makes detection far more reliable.
- If the table is part of a scan, OCR will get you the words, but the grid has to be rebuilt by hand or with Convert Text to Table. Tesseract’s own documentation notes it has trouble recognising text in tables.
- If the original spreadsheet exists, ask for it. A PDF of a spreadsheet is always a step down from the spreadsheet.
Frequently asked questions
Why did my PDF table turn into plain text in Word?
Converters find tables mostly from the lines drawn between cells. Tables without visible lines, and any table in a scanned PDF, are much more likely to come out as text with tabs or spaces between the values.
How do I turn tab-separated text back into a table in Word?
Select the lines, then choose Insert, Table, Convert Text to Table, and pick Tabs as the separator.