Extract Tables from PDF
Find the tables of a PDF and save them as JSON, Markdown or CSV. A table is found by its columns, with or without ruled lines, and Arabic is written as it is read. Check the result. Runs in your browser.
About This Tool
Extract Tables from PDF finds the tables in a PDF and saves them as JSON, Markdown or CSV, all the tables in one file. A table is lines of text that keep to the same columns: the tool looks at where the type lies on the page, so a table is found whether it is drawn with ruled lines or laid out with spacing alone, and whichever edge its columns are set against. Text that is not in a table is left out. It needs a PDF with real text: in a scan the words are pictures, so run OCR PDF on it first.
Each table keeps its columns where a cell is empty, and an Arabic table is written from its right-hand column, each cell in reading order. Before you download, the page lists the tables it found, with the page each is on and its first rows. This is a reading of the page, so check it against the PDF: a cell whose text runs over two lines becomes two rows, a table of long sentences can be taken for columns of text and left out, and a list set in two short columns can be taken for a table. Use "Fewest columns" and "Fewest rows" to leave out tables smaller than you want.
Cells are written as the page has them. In CSV, a cell that a spreadsheet would run as a formula (one that begins with = or @, or with + or - before anything but a figure) is given a leading apostrophe so that it stays text. The file is made in your browser and is not uploaded.
How to Use
Upload Your PDF
Upload the PDF containing tables you want to extract.
Configure Detection
Choose JSON, Markdown or CSV. Under "Pages", type the pages to look at, such as 1-5, 8, or leave it empty for all. "Fewest columns" and "Fewest rows" leave out tables smaller than that.
Export and Download
Click "Extract Tables". The page says how many tables were found and shows the first rows of each; click "Download" to save the file. With "Say which page each table is on" ticked, each table is headed with its number and its page.
Use Cases
Data Analysis
Extract table data for analysis in spreadsheets or databases.
Report Processing
Pull tables from PDF reports for further processing.
Documentation
Convert PDF tables to Markdown for technical documentation.
Frequently Asked Questions
Can it detect complex tables?
It finds a table by its columns: lines of text with clear paper between them at the same places. Ruled lines, borders and shading are not looked at, and merged cells are not known as such. A cell that runs over two lines becomes two rows, and a table whose cells are long sentences may be taken for columns of prose and left out.
What if no tables are found?
You are told so, and no file is made. If the PDF is a scan, the message says that it has no text: run OCR PDF first. If tables were found but all were smaller than "Fewest columns" and "Fewest rows" ask for, the message says how many, and lowering those two brings them in. To get all the text of a page and not only its tables, use PDF to Excel or PDF to Markdown.
Can I extract from specific pages?
Yes. Type them under "Pages", for example 1-5, 8. If nothing you type matches a page of the file, you are told so and nothing is extracted.