Your PDF has a beautiful table — except the top row has a header that spans three columns, and another that spans two. You convert it, open in Excel, and the headers have collapsed into the leftmost column. Two of the three are empty.
Merged cells. The hardest single thing in PDF-to-Excel.
Why merges are hard to detect
A PDF stores no concept of "merge." It stores a piece of text at a specific position. When a converter sees text that spans the visual width of three columns, it has to decide: is this a merged header for those three columns, or just text that happens to be wide?
Good converters get this right most of the time. Borderless tables make it harder — they have less geometric signal to work with.
What Flint does about it
Flint's PDF to Excel attempts to detect merge zones based on column boundaries, alignment and surrounding rows. Headers that clearly span multiple columns usually come across as merged. Borderless variants are detected if the spacing pattern is consistent.
If you spot a merged header collapsed to one cell after conversion, manually select the cells and Excel > Format > Merge & Center. Five seconds per header.
Vertical merges (row spans)
Cells that span multiple rows — common in matrix-style tables — are harder again. The converter often duplicates the value across the rows it should span, which is actually more useful for filtering and pivoting than a true merge. If you genuinely need a visual merge, manually merge after.
Most analytical workflows prefer the duplicated values anyway: pivot tables, charts and SUMIFs all behave better with explicit values than with sparse merged cells.
When merges define the table's meaning
Some tables are basically unreadable without the merges (e.g. "this header applies to columns 2–4"). In those cases, after conversion, add a row above the column headers labelling the meta-group. Or, if it matters for visual reporting, manually re-merge the relevant cells. For analysis, leave them unmerged.
FAQ
Are vertical merges supported?
Detection is improving but less reliable than horizontal. The converter often duplicates values down rows instead, which is usually more useful for analysis.
Can I tell the converter where the merges are?
Not in the free flow. The converter infers automatically. Manual cleanup in Excel is the practical answer for unusual layouts.
Will pivot tables work afterwards?
Yes — and they'll behave better if cells are unmerged with duplicated values rather than visually merged. Pivots dislike sparse merged cells.
Why are some merged headers split into multiple cells?
The converter saw enough horizontal spacing to treat them as separate columns. Easy to manually merge in Excel.
Merges are imperfect but mostly recoverable. Convert your PDF to Excel and budget a minute for cleanup.