Convert PDF to CSV

What the table detector needs from the PDF

A PDF stores words at coordinates on a page. It has no idea of a cell. PDF to CSV looks at the native text layer, finds words that line up in repeated rows and columns, and treats each aligned block as a table. That is an inference from position, not a read of a structure, so it works on the tables a PDF usually holds: an export from a database, an invoice line list, a price list, a statement, a report table with ruled lines or clear column gaps.

Three things defeat it. A scanned PDF has no text layer, only a picture of a table, and the detector has nothing to align. The tool refuses it instead of producing an empty file, and the fix is text recognition first. A table whose columns are separated by a single space, so that the gap inside a cell matches the gap between cells, splits in the wrong places. A page laid out in two text columns, as in a journal article, looks like a two-column table to a position-based detector and has to be deselected.

Merged cells are a milder problem. A heading that spans two columns lands in one of them, and the other is empty, because CSV has no way to say that a cell is wide. Rotated tables and forms with a box drawn around each single value are usually not detected at all.

How to convert PDF to CSV file by file

Drop the PDF on the page. The tool lists the tables it found, each with the page it sits on. Select the ones you want. One selection downloads as a single .csv. Several download as a ZIP holding one CSV per table, and the tool never joins unrelated tables into one file. A statement that runs over four pages therefore arrives as four CSVs, and you join them in the spreadsheet, where you can see the seam.

Then choose how cells are written. In spreadsheet-safe mode a cell that begins with =, +, - or @ gets a leading apostrophe, so that a CSV made from an untrusted PDF cannot run a formula when it is opened. In exact-text mode the cell is written as it was read, which is what you want when the values are negative numbers with a leading minus. Every cell that holds a comma, a quote or a line break is quoted, and a quote inside a value is doubled, which is the form every spreadsheet expects.

Everything happens in this browser tab: the PDF, the detected cells, the preview and the CSV. Nothing is uploaded, and there is no account.

What CSV cannot hold, and when Excel is the better output

CSV is values separated by commas. Bold, colour, borders, column width, number formats, merged cells, a second sheet: none of it survives, because there is nowhere to put it. That is a feature when the CSV is going into a database, a script or an import that wants plain values. It is a loss when the table is going back to a person.

PDF to Excel uses the same detector and writes a workbook instead, with each detected table on its own sheet, and reads the workbook back before the download. Choose it when the PDF holds several tables you want in one file, when the values should stay numbers, or when someone is going to open the result in Excel and read it. Convert PDF to Excel covers that route in full. Choose CSV when the next step is code, a database load, a mail merge or any tool that reads plain text.

How to check a PDF table to CSV result

The tool's own check proves one thing: that the CSV parses back into the same rows, columns and cell strings the detector produced. It cannot prove that the detector drew the column boundaries where the author of the PDF meant them. That part is yours, and it takes a minute.

1. Count the rows. A statement page says how many transactions it holds, or you count them. The CSV, opened in a spreadsheet, must have that many rows plus a header. A short count means a row merged with its neighbour. A long one means a wrapped description split into two rows.
2. Sum a column. Add the amount column and compare it with the total printed on the PDF. A match is strong evidence that every value landed in the right column with the right sign.
3. Read the first and last row against the preview. The tool shows the source page beside the detected cells, and a column that shifted by one shows up there before it shows up in a spreadsheet.
4. Look for cells that hold two values. A description that ran into the amount column reads as text where a number should be, and a spreadsheet sorts it to the wrong end.

Convert a bank statement PDF to Excel walks through the same checks on the most common case, a statement whose descriptions wrap to a second line. The rest of the spreadsheet and document guides are under PDF, Word and Excel help.