PDF → Csv

Beta

Convert one table in a PDF to CSV by drawing around it, setting the columns and checking each cell.

How to convert a PDF table to CSV

Choose a PDF and the page that holds the table.
Draw one region around the table on the page preview.
Set the column direction and divider positions, then press Read selected region.
Check every cell in the grid, edit any that are wrong, and add or remove rows.
Choose how formula-like cells are handled, then press Export reviewed CSV.

Options

Table page and region: Pick one page and draw one rectangle around the table. Changing the page or the region clears the grid. Clear region removes the rectangle.
Column direction: Across the page (left to right) or down the page (top to bottom), as the page is displayed. Rotated pages start with down the page.
Divider positions: Comma-separated percentages across the region, for example 35, 70 for three columns. Leave it empty for a single column. Up to 12 columns.
Read selected region: Reads the selectable text inside the region into an editable grid. Run it again after changing columns. That replaces your cell edits.
Grid editing: Every cell is editable. Add row and Remove fix rows that were split or merged by wrapped text.
Formula-leading cells: Spreadsheet-safe adds an apostrophe before values starting with =, +, - or @ so Excel and similar apps do not run them as formulas. Literal CSV keeps every value exactly as reviewed.

What you get

One UTF-8 CSV file of the grid you reviewed, with standard quoting and CRLF line endings.
Only the one region on one page is exported. Other tables and text are left out.
Formatting, merged cells and numbers as numeric types are not kept. Every cell is text.

Limits

PDFs up to 32 MB and 500 pages. One table region on one page per export.
Up to 200 rows, 12 columns, 1,200 cells and 256,000 characters of grid text. The CSV can be up to 2 MB.
No OCR. Scanned tables need Make PDF searchable first.
Columns are not detected automatically. You set the dividers.
Wrapped rows, merged cells and right-to-left text can land in the wrong cell, so check each one.
Excel (XLSX) export is not available. Open the CSV in your spreadsheet app instead.

Check the result

Deckle reopens the saved CSV and checks that it matches the grid you reviewed, cell for cell. That check does not compare the grid with the PDF, so look over the cells against the page yourself, including any apostrophe-prefixed values.

Questions

Can it find tables in my PDF automatically?

No. You draw around the table and place the column dividers. The tool then sorts the text inside into rows and columns for you to check.

Can I export several tables at once?

No. Each export covers one region on one page. Run it again for each table and combine the CSV files in your spreadsheet.

Why do some cells start with an apostrophe?

Spreadsheet-safe mode adds it to values that begin with =, +, - or @, so a spreadsheet does not run them as formulas. Choose Literal CSV if you need the exact values, such as negative numbers.

Does it work on scanned tables?

Not on its own. A scan has no selectable text, so the region reads as empty. Make the PDF searchable first, then extract from the searchable copy and check every cell.

Can I get an Excel file?

Not yet. The tool exports CSV, which Excel, Numbers, LibreOffice and Google Sheets all open.

Is the PDF uploaded?

No. The text is read and the CSV is written in your browser.