Why does copying a PDF table lose the columns?

A PDF stores where each piece of text is drawn, not that it belongs to a column. There are no cells in the file. Anything that recovers a table is inferring it from the position of the text and the ruling lines, which is why results vary so much between tools.

How do I know if my PDF is scanned?

Try to select the text with your cursor. If nothing highlights, the page is a picture of a document and there is no text in it to extract. A scan needs OCR before any table can be read.

Is there a free way to do this?

Yes. Tabula is free, open source and runs on your own machine, and it is a good answer for a text-native PDF if you are willing to install a Java runtime. It cannot open a scan.

How to extract tables from a PDF

Four methods work, and which one you need depends on a single fact about your file that nobody tells you to check first.

First: does your PDF have a text layer?

Open the file and try to select a row of the table with your cursor.

The text highlights — it is text-native. Everything below works.
Nothing highlights — the page is a picture. There is no text in the file at all, and half the methods below cannot work on it.

This is why people conclude that “PDF table tools do not work”: they tried a rules-based tool on a scan.

The four methods

1. Copy and paste

For one small table in a text-native PDF, select it and paste into a spreadsheet. The columns usually collapse, but for ten rows that is faster than installing anything. Try this first.

2. Tabula, free and on your own machine

Tabula is open source, uploads nothing, and is genuinely good on text-native PDFs. It wants a Java runtime, and it cannot open a scan at all.

3. OCR, if the page is an image

A scan has to be read as an image before any table exists to extract: a different and more expensive operation, which is why a scanned bank statement is harder than a digital one.

4. An API, for many files or on a schedule

Also the answer for a scan, and for a table that runs across a page break.

Check your own file, free, before choosing

Free, no account. If it reports no tables on a page you can see one on, that page is a scan.

Which pages of your PDF have tables?

Free, no signup, no credit charged. Tells you what extraction would cost before you buy anything.

Choose a PDF or drop it hereTells you what extraction would cost, free.

This one sends your file. Finding tables means reading the PDF’s rulings and glyph positions, which is not something a browser does. It is scanned in memory and deleted when the request ends, and nothing is extracted or stored.

Charged per table-bearing page, $0.09 each on the $14.99 pack: a forty-page report with four table pages costs four. More about the extractor · every price.