Why Is My PDF Not Converting to Excel Properly? 9 Causes and Fixes

Last updated July 2026.

A PDF converts to Excel badly because the PDF does not contain a table. It contains characters positioned at coordinates on a page. Any converter has to guess where the columns and rows were, and on documents without gridlines, with wrapped text, or with repeated headers, those guesses fail in predictable ways. Below are the nine failures people actually hit, what causes each one, and how to fix it.

The root cause, in one paragraph

When a bank, an accounting system, or a report generator writes a PDF, it does not save a spreadsheet. It saves instructions like "draw the text 1,240.55 at x=487, y=612 in 9pt Helvetica." The visual columns you see are an illusion produced by consistent x coordinates. Nothing in the file says "this is column three of a table." So when a converter opens that PDF it is doing reverse engineering: clustering characters by position, deciding where one row stops and the next begins, and hoping the document was laid out consistently. Most conversion problems are that inference going wrong, not a bug in your software.

1. Everything lands in one column

You open the XLSX and every line of the page is jammed into column A. This means the converter fell back to plain text extraction because it could not detect a table structure at all. It happens most often with PDFs that have no ruling lines and irregular spacing, and with files where the text layer was produced by a poor OCR pass.

Fix: use a tool that does table detection rather than text extraction, or in Excel use Data, then Text to Columns, and split on fixed width if the original was monospaced. For anything with variable-width descriptions, fixed width will not save you and you need a converter that understands the document type.

2. Merged cells everywhere

Merged cells are a layout-faithful exporter trying to reproduce how the page looked. When a heading spans the width of the page, or a description sits between two column boundaries, the exporter merges cells to keep the visual position. The result looks like the PDF and is unusable for sorting or pivoting.

Fix: select all, then Home, Merge and Center, Unmerge Cells. That leaves the value in the top-left cell of each former merge and blanks the rest, so you will still need to fill down. The better fix is to convert with a tool that outputs data rather than layout, because unmerging a 400-row export by hand introduces its own errors.

3. One row split into two or three

This is the most damaging failure because it silently changes your row count. It happens when a long description wraps onto a second printed line. The converter sees two lines of text at two different y coordinates and creates two rows, one of which has a date and an amount and one of which has only a fragment of text.

Fix: sort or filter on the amount column and look for rows with no value. Those orphan rows are almost always continuation text belonging to the row above. Rejoining them manually is fine on a short file and dangerous on a long one, because attaching a description to the wrong transaction is an error that survives review. Converters built for a specific document type handle wrapping as a rule rather than a guess.

4. Numbers arrive as text and will not sum

You highlight the amount column, look at the status bar, and the sum is blank or zero. Excel is treating the values as text. The usual culprits are parentheses used for negatives, a trailing CR or DR marker, a currency symbol, a non-breaking space as the thousands separator, or a minus sign that is actually a Unicode dash character rather than a hyphen.

Fix: select the column, then Data, Text to Columns, Next, Next, and set the column format to General before finishing. That forces re-evaluation of every cell. If parentheses are the issue, find and replace the opening parenthesis with a minus sign and delete the closing one first. If a stray character is the problem, use CLEAN and TRIM together, or SUBSTITUTE with CHAR(160) to strip non-breaking spaces.

5. Dates flip between formats or become serial numbers

Some rows read 03/04/2026 and some read 4-Mar, or an entire column turns into five-digit numbers. Excel interprets each imported value against your regional settings, and a US MM/DD/YYYY file opened on a machine set to DD/MM/YYYY will silently swap day and month on every date where both are 12 or lower.

Fix: import through Power Query rather than opening the file directly. Use Data, Get Data, From File, and when you set the date column type choose Using Locale so you can tell Excel the source format explicitly. This is the only reliable way to stop the swap, and it matters enormously on financial data where a March 4 transaction becoming April 3 changes which period it falls in.

6. Columns are shifted or interleaved

Amounts appear under the description header, or the balance column contains a mix of balances and transaction amounts. On a digital PDF this usually means the column boundaries were inconsistent between pages. On a scanned document it usually means the page was slightly rotated when it was scanned, so the x coordinates drift as you move down the page.

Fix: for a scan, straighten the source. Rescan at 300 DPI with the page square to the glass, or use a deskew step before OCR. Even a two degree skew is enough to push the rightmost column off its expected position by the bottom of the page. For digital PDFs, a converter tuned to the specific document layout handles the inconsistency because it knows what the columns are supposed to be.

7. Page headers, footers, and marketing blocks are mixed into the data

Every few dozen rows you find the account holder name, the statement period, a page number, or a promotional panel sitting in the middle of the transactions. Nothing is wrong: the converter faithfully exported everything on the page, and those elements were on every page.

Fix: filter the data on a column that only real rows have, such as the amount, then delete the visible non-data rows in one pass. Do not sort first, because sorting a table that still contains header fragments scrambles the transaction order you may need later. Converters built for one document type strip repeated furniture automatically.

8. A scanned PDF produces nothing at all

The export is empty or contains a few garbage characters. The PDF is an image with no text layer, which happens with anything scanned, photographed, or faxed. There is nothing for a text extractor to extract.

Fix: run OCR first, then convert. Quality of the source matters more than the OCR engine: scan at 300 DPI in grayscale rather than black and white, keep the page flat, and avoid photographing a statement at an angle. A photo taken on a desk under overhead lighting, slightly skewed, is the single most common reason OCR output is unusable, and no amount of software compensates for it fully.

9. Multi-page documents lose rows at the page breaks

The first page converts perfectly, then a handful of rows go missing around each page boundary. This happens when a converter processes pages independently and drops rows that straddle a break, or discards the first rows of a page because it mistook them for header content.

Fix: check the row count of the output against the source before anything else. This is the check that catches the failure that costs the most, because a missing row is invisible in the spreadsheet. On a financial document you have a built-in test: opening balance plus the sum of all amounts should equal the closing balance printed at the end. If it does not tie, rows are missing.

The three checks that catch every one of these

Whatever tool you use, run the same short test on the output before you build anything on it.

• Row count. Count transaction rows in the spreadsheet and compare with the source document. A mismatch means split rows or dropped rows.
• Footing. On a financial document, opening balance plus the sum of amounts must equal the closing balance. This catches missing rows that a row count alone would miss if a split also occurred.
• Sign consistency. Look at debits and credits on the first page and the last page. If the convention changed halfway through, every total downstream is wrong.

Ninety seconds of checking saves the hour you would otherwise spend at reconciliation working out which of four hundred rows is wrong.

Why bank statements are the hardest case

Almost every failure above shows up worse on a bank statement than on any other business document. Statements have no gridlines, so column boundaries exist only as whitespace. Merchant descriptions wrap constantly. Debits and credits appear in separate columns at one bank, in one column with parentheses at another, and with DR and CR suffixes at a third. The account header and balance summary repeat on every page. Some issuers insert a promotional block mid-list. And the document ends with a closing balance that publicly grades your extraction: either the rows add up to it or they do not.

That is why a general PDF to Excel exporter and a converter tuned to statement layouts behave so differently on the same file. The general tool asks what is on this page. The tuned tool asks which of this is a transaction, and it already knows which banks wrap descriptions and which repeat their summary block. If statements are what you are converting, our bank statement PDF to Excel converter handles the wrapping, header stripping, and sign normalization as rules rather than guesses, and keeps the running balance as its own column so the footing check is available to you.

When the manual route is still the right call

If you have one short document and it is a well-ruled table, importing it directly through Power Query is genuinely the fastest path and costs nothing. Data, Get Data, From File, From PDF lets Excel list the tables it detected and you pick the one you want, with type and locale set properly on the way in. Our walkthrough on how to import a bank statement PDF into Excel with Power Query covers the steps, including the date locale setting that stops the day and month swapping.

The calculation changes with volume. One document is a ten minute job. Nine months of catch-up across three accounts is thirty to forty documents, each with its own cleanup pass, and the failure modes above compound: a split row here, a text amount there, a missing page-break transaction that nobody notices until the books will not reconcile. At that point the cleanup is the work, not a preliminary to it.

What to do once the data is clean

A clean export is the starting point, not the deliverable. From properly typed columns you can categorize transactions from a bank statement into expense categories, reconcile against your books, or build a profit and loss report for the period. If the documents you are fighting with are invoices rather than statements, the same layout problems apply and a tool that turns them into invoice data in a spreadsheet saves the same cleanup pass.

Frequently asked questions

Why is my PDF not converting to Excel properly?

Because a PDF stores characters at page coordinates, not a table. The converter has to infer where rows and columns were, and that inference fails on documents without gridlines, with wrapped descriptions, or with repeated page headers. The result is merged cells, split rows, and amounts stored as text.

How do I convert a PDF to Excel without losing formatting?

You generally do not want the formatting, you want the data. Aim for an output where each transaction is one row and each field is one properly typed column. Preserving the visual layout is exactly what produces merged cells and blank spacer rows, which is the problem rather than the goal.

Why do my numbers not add up after converting a PDF to Excel?

The values are stored as text. Parentheses for negatives, trailing CR or DR markers, currency symbols, and non-breaking spaces all stop Excel treating a cell as numeric. Run the column through Data, Text to Columns and set the format to General to force re-evaluation.

How do I fix merged cells after a PDF to Excel conversion?

Select all, then Home, Merge and Center, Unmerge Cells. Values remain only in the top-left cell of each former merge, so fill down where needed. On large files this is error-prone, and converting with a tool that outputs data rather than layout avoids the step entirely.

Why does my scanned PDF convert to an empty Excel file?

A scan is an image with no text layer, so there is nothing to extract. Run OCR first. Scan quality matters more than the engine: 300 DPI, grayscale, page square to the glass. A skewed phone photo is the most common reason OCR output stays unusable.

Is Power Query better than a PDF converter for Excel?

Power Query is excellent and free for well-structured tables, and its Using Locale option is the correct fix for date swapping. It struggles with the same things every generic tool struggles with: no gridlines, wrapped descriptions, and repeated page headers, which is precisely what bank statements have.

How do I know if rows went missing during conversion?

Compare the row count in the output with the source, then check the arithmetic. On a financial document, opening balance plus the sum of all amounts should equal the closing balance printed at the end. If it does not tie, rows are missing or duplicated, and no visual inspection will find them faster.