Why does copying a table from a PDF into Excel put everything in one column?

Because a PDF has no real table — it's text drawn at coordinates to look like a grid. Copy-paste grabs the text in reading order and drops it into one cell, since there's no row-and-column structure to copy. Getting a real grid requires a tool that reconstructs the columns from where the characters sit.

Can I extract a table from a scanned PDF?

Not directly — a scanned PDF is a picture of a table, with no text to extract. Run it through OCR first to turn the image into real text, then extract the table from the OCR'd version.

Why do my extracted columns keep merging together?

Usually because the PDF has whitespace-only characters positioned inside the gaps between columns, which fools the column detector into treating two columns as one. A tool that strips those phantom characters before detecting columns avoids it. It's a common failure that looks like the tool can't read tables at all.

Will the extracted numbers work as numbers in Excel?

They arrive as text, and Excel may not treat them as numbers until you convert the column — currency symbols and thousands separators come across as part of the string. Select the column and use Excel's convert-to-number to make them calculable.

How to extract a table from a PDF into Excel

Why copy-paste always mangles it

Select a table in a PDF, copy it, paste into Excel, and you get a single column of jumbled text where a grid used to be. This isn't Excel being stupid — it's the PDF being honest about what it actually is. A spreadsheet stores a grid: this value is in row 4, column C, full stop. A PDF stores instructions to draw characters at particular coordinates on a page. When you look at a PDF table, the grid exists only in your eyes; the file has no rows, no columns and no cells.

So copy-paste has nothing structural to copy. It grabs the text in reading order and drops it into one cell, and the neat alignment you saw — which was just ink positioned to look aligned — is gone. Getting the table into Excel means reconstructing the grid from where the characters landed, which is a genuinely harder job than it looks.

What extraction can recover — and what it can't

Recovers well: machine-generated tables with consistent column alignment — the great majority of statements and reports.
Struggles with: merged header cells (a header spanning two columns has to be assigned to one, and that's a guess), and tables with wildly uneven spacing.
Can't recover anything from a scanned PDF: that's a picture of a table, not text. Nothing can extract data from an image without OCR first.
Splits multi-page tables page by page — you stitch them back together in the spreadsheet.

Reconstruction works well when the PDF was generated by software. Bank statements, invoices, exported reports — their columns line up to the pixel because a program placed them, so the boundaries between columns are clean and easy to find. It works less well on tables that were scanned, hand-laid-out, or built with uneven spacing, because the geometry it has to read is genuinely ambiguous.

The invisible-character problem

Here's a failure that makes extraction look broken when it isn't. Columns are found by looking for the vertical gaps between them — a corridor of blank space is a column boundary. Simple, and it fails on a lot of real PDFs, because PDF renderers sometimes emit whitespace-only text sitting inside those gaps. A space character with no visible glyph, positioned exactly where the corridor should be, silently bridges two columns into one, and suddenly your amounts and your dates are in the same cell.

A good extractor strips those phantom items before measuring the gaps, which is why a statement whose columns kept merging in one tool comes apart correctly in another. If your extractions keep merging columns, this is usually why — it's not that the tool can't read tables, it's that it's reading a space that isn't really there.

Getting clean data out

Check whether the PDF is text or a scan first: try to select text in it. If you can't, it's a scan — run it through [OCR](/utility/ocr/) to get real text, then extract.
Expect values as text. Amounts come across as amounts, but Excel may not treat them as numbers until you convert the column, because currency symbols and separators are part of the extracted string.
Verify the header row and any merged cells by hand — they're where the guessing happens.
If whoever sent the PDF can send the original .xlsx or CSV, ask for it. That file is exact; an extraction is inferred. Extraction is for when the data exists only as a PDF.

SnapToolz's [PDF to Excel](/pdf/to-excel/) strips the phantom whitespace items before detecting columns, which is what keeps a bank statement's columns from merging. For the reverse trip — a spreadsheet you want as a clean printable report — [Excel to PDF](/pdf/from-excel/) reads the workbook's own number formats so currency stays currency and dates stay dates.

Tools used in this guide

FAQ

Why does copying a table from a PDF into Excel put everything in one column?: Because a PDF has no real table — it's text drawn at coordinates to look like a grid. Copy-paste grabs the text in reading order and drops it into one cell, since there's no row-and-column structure to copy. Getting a real grid requires a tool that reconstructs the columns from where the characters sit.
Can I extract a table from a scanned PDF?: Not directly — a scanned PDF is a picture of a table, with no text to extract. Run it through OCR first to turn the image into real text, then extract the table from the OCR'd version.
Why do my extracted columns keep merging together?: Usually because the PDF has whitespace-only characters positioned inside the gaps between columns, which fools the column detector into treating two columns as one. A tool that strips those phantom characters before detecting columns avoids it. It's a common failure that looks like the tool can't read tables at all.
Will the extracted numbers work as numbers in Excel?: They arrive as text, and Excel may not treat them as numbers until you convert the column — currency symbols and thousands separators come across as part of the string. Select the column and use Excel's convert-to-number to make them calculable.