Skip to main content
Datatool360

How to Convert a PDF to Text (Columns, Tables and Scans)

By the Datatool360 team ·

You select a page of a PDF, paste it somewhere else, and get broken lines, two columns mixed into one and a page number in the middle of a sentence. It is not your fault and not the fault of your computer. This guide explains what is going on and how to get text that you can actually use.

Want to do it right now? Open the PDF to Text tool

Why copy and paste goes wrong

A PDF does not remember that two columns belong to one article. It only knows where each piece of text sits on the page. A program that reads the pieces from top to bottom therefore jumps between the left and the right column.

The same thing causes the other problems. Tables lose their cells and become a pile of numbers. Every line ends with a line break, even in the middle of a sentence. Words are split by a hyphen where the line ended. And the header, the footer and the page number sit in the middle of the text, because they are on the page too.

Convert a PDF to text step by step

  1. Open the PDF to Text tool and upload your PDF.
  2. Wait while the text is read. Pages that are pictures are read with OCR, which takes longer.
  3. Compare the text with the original page, which is shown next to it, and correct what you need in the editor.
  4. Copy the text, or download it as TXT, Markdown, Word or JSON.

What the converter does for you

  • Columns are read one after the other, so a two column article comes out in the right order.
  • Tables are found and written as aligned text, as Markdown tables, or as real tables in the Word file.
  • Headings, lists and paragraphs are recognised. Lines that were only wrapped are joined again, including words that were split by a hyphen.
  • Headers, footers and page numbers that repeat on every page are taken out.

Which format to choose? TXT if you only need the words. Markdown if you want headings and tables that you can paste into notes or a website. Word if you want to keep working on it as a document. JSON if the text goes into a program.

Scanned documents

A scan has no text at all, only a picture of it. So the tool reads the picture with OCR (English). The quality depends heavily on the scan. A straight page, a resolution of about 300 dpi and good contrast give the best results. A skewed, faded or photographed page gives more mistakes, and handwriting is not supported.

If you can, scan again with better settings instead of fixing a bad result by hand. And always check numbers, dates and names: OCR confuses 0 and O, 1 and l, and 5 and S more often than you would think.

Limits to know about

  • Very irregular tables, for example with merged cells, may come out as simple rows. Check the important ones against the page.
  • The file can be up to 50 MB and 300 pages.
  • Unlike our editing tools, this converter does send your PDF to our server. It is kept in memory for about 30 minutes so that you can view and download the result, and it is not written to disk.
  • For equations, use the Math PDF converter instead. It keeps them as LaTeX.

Keep reading