When a PDF table doesn't convert cleanly to Excel — columns shift, rows split, or cells merge — it's usually because the PDF's table has no real structure for the converter to read. A few adjustments to how you convert, plus some quick cleanup, gets you a usable spreadsheet.
Why PDF tables convert badly
PDFs don't store tables as tables — they store text at fixed positions that happen to look like a grid. A converter has to guess where the columns and rows are. When borders are missing, cells are merged, or a row wraps onto two lines, that guess can go wrong, and the spreadsheet comes out misaligned.
Common problems and fixes
- Columns merged into one → the source had no clear separators; try CSV output and split the column in Excel (Data → Text to Columns).
- A row split across two → wrapped text; merge the two rows back together.
- Extra header or footer rows → just delete them after converting.
- Numbers imported as text → select the column and use Excel's "Convert to Number."
Try CSV instead of Excel
If the .xlsx output is messy, convert to CSV instead. Raw comma-separated values are simpler and often easier to clean up or re-import than a mis-structured spreadsheet.
If it's a scanned PDF
OCR-based extraction is less precise than reading digital text, so a scanned table will need more cleanup. Use the clearest scan you have and expect to correct a few figures — see converting a scanned invoice to Excel.
When to just clean it by hand
For a one-off small table, converting and then tidying in Excel is faster than fighting the layout. For a repeating format like a monthly statement, it's worth finding the output — Excel or CSV — that converts most cleanly and sticking with it. Both PDF to Excel and PDF to CSV run on our secure server (uploaded, converted, deleted after, and labeled as server tools) — try both and keep whichever comes out cleaner.
Why a table that looks like a table is not one
A PDF has no concept of a table. It positions text at coordinates, and what you perceive as rows and columns is an arrangement your eye assembles. Extraction works by inferring structure from that positioning, which is why it succeeds on a ruled, evenly spaced financial table and struggles on one held together by whitespace alone.
The layouts that reliably break it
- Merged cells and multi-row headers, which have no single unambiguous grid to map onto
- Rows that wrap onto a second line, which look like separate records
- Tables split across a page break, where the header does not repeat
- Columns separated only by spacing, with no rules and inconsistent gaps
If extraction produces a misaligned sheet, check whether the PDF is scanned at all: if your cursor draws a box instead of selecting text, there is no text to extract and OCR has to come first. Where the table is genuinely irregular, pulling plain text and reshaping it in the spreadsheet is often faster than fighting the layout.