To get a PDF table into Excel with its columns intact, use a converter that finds the column boundaries for you, such as PDF to Excel in a browser tab or, in Excel for Microsoft 365 on Windows, Data > Get Data > From File > From PDF. Then check the row count and the totals against the PDF. Columns usually survive when the PDF draws lines between them or keeps each column in the same place on every row. They break on merged cells and on designed layouts where only uneven spacing suggests the columns, because a PDF has no table in it to convert.
A PDF has no tables
Open a spreadsheet file and it contains cells: a grid, with a value in each position. The structure is the data.
A PDF page contains none of that. It contains instructions like draw the characters “4,182.60” at x=412, y=308 in this font, and separately, perhaps, draw a line from here to there. A human looking at the result sees a table. The file contains text at coordinates and some lines that happen to sit between them.
Every PDF-to-Excel converter therefore does the same thing: it infers the grid. It looks for ruling lines, or for text that aligns into vertical bands across several rows, and decides where the cells must be. On a clean financial table that inference is nearly always right. On a designed layout it is nearly always partly wrong.
Knowing which one you have tells you how much checking to do.
What converts well, and what does not
Converts cleanly:
- Ruled tables — lines around cells or at least between columns
- Consistent column positions down the whole table
- One line of text per cell
- Numbers right-aligned in their columns, which makes boundaries obvious
Converts partially:
- Tables with no ruling lines but strong, even spacing
- Tables that continue across pages with repeated headers
- Cells containing wrapped text over two or three lines
PDF to Excel puts a table that runs off the foot of one page and onto the top of the next back together in one sheet when its columns sit in the same places on both pages, and drops the header row if it is repeated.
Does not convert reliably, and no tool will change that:
- Merged cells spanning columns or rows
- Nested tables
- Layouts where whitespace alone implies structure — a magazine price list, a designed report page
- Scans with no text layer, until OCR has been run
That last category is genuinely hard, not a gap in one product. Pulling structure out of an unruled layout needs a document-layout model far too large to send to a browser tab, which is why Filozy says so plainly instead of pretending.
Step by step
- Check for a text layer. Try selecting a number in the PDF. If nothing selects, it is a scan: run OCR PDF first.
- Open PDF to Excel and choose the file. It is read in your browser.
- Read the result summary. It says how many tables and rows it found and in how many sheets, and the notes under it name any page whose grid was left out because its columns did not line up well enough to be sure. A row count far from what you expect is the signal to stop and check.
- Open the spreadsheet and verify in this order:
- Row count against the PDF. Wrong here means the grid was misread; everything downstream is suspect.
- Column boundaries on the widest row, where misalignment shows first.
- Totals. If the PDF has a sum, re-sum the column in the spreadsheet and compare. This single check catches most silent errors.
- Decimal points and negatives. PDF to Excel reads (2,100) as −2,100 and says so in its notes. Check that a bracketed figure that was never negative, such as a note reference (3), has not become −3. If you used a different converter, check that bracketed figures are numbers, not text: SUM leaves text cells out of the total without an error.
- Keep the PDF. The spreadsheet is a reconstruction, and the PDF stays the source of record.
If you have Excel for Microsoft 365 on Windows
Excel can read a PDF’s tables itself: Data > Get Data > From File > From PDF. Choose the file, pick a table in the Navigator, then select Load, or Transform Data to fix headers and split columns first. Like any converter, it infers the grid, so the checks above still apply. Microsoft documents the connector in its Power Query PDF connector reference, including the option that joins similar tables on consecutive pages. Without that version of Excel, use PDF to Excel: it reads the file in your browser tab and leaves out any grid whose columns do not line up well enough to be sure it is a table.
When a spreadsheet is the wrong target
If you want the words rather than the figures, PDF to Markdown gives you the text as Markdown without forcing anything into cells that have to be numerically correct. It reads each line straight across the page, so a table comes out as running text, not a grid.
And if the table originated in a spreadsheet that still exists somewhere, ask for that file. No conversion beats the original.