A landscape PDF converts to Excel cleanly when the tool reads the page at its real width and treats each wrapped header as one label. Check whether the page is truly wide or a rotated portrait page, count the columns before and after, and handle any portrait pages in the same file as a separate table.
Why a landscape PDF table loses columns in Excel
A landscape PDF table loses columns when the converter misjudges where one column ends and the next begins. Wide tables pack many narrow columns close together, headers wrap onto two or three lines, and small gaps between figures look like spaces inside one value. Any of these can merge two columns or push values into the wrong one.
A PDF has no columns of its own. It stores pieces of text at positions on the page, and the converter guesses the column boundaries from the gaps between them. On a portrait statement with five columns, the gaps are wide and the guess is easy. On a landscape report with twelve columns, the gaps shrink. A long description can run up to the edge of the next column, and an amount with a minus sign can sit right against its neighbour.
Three things go wrong most often:
- Two columns become one. A "Fees" column and a "Tax" column with small figures and a narrow gap come out as a single cell holding "2.50 0.38".
- A column is dropped at the edge. Some tools read a fixed portrait width and cut off everything past it. The right-hand columns, usually the totals and balances, are the ones that vanish.
- Values shift one column left. If a row has an empty cell in the middle, the next value slides into the gap, and every column to the right of it is off by one.
The edge case is easy to catch. Count the columns in the PDF, count them in the sheet, and compare. If the sheet is short, look at the right-hand side first. If the count matches but figures look wrong, look for one row where a blank cell let the values slide.
Is the page really landscape, or a rotated portrait page?
A page can look landscape for two different reasons. It may have been laid out wide from the start, or it may be a portrait page that is only displayed sideways. The difference matters because a sideways page stores its text running up the page, which many converters read as one letter per line.
The PDF 32000-1:2008 specification explains how the sideways case happens. Each page can carry a Rotate entry, defined in the page object as "the number of degrees by which the page shall be rotated clockwise when displayed or printed", always a multiple of 90. The viewer turns the page for you, so it looks wide on screen. Underneath, the content may still be drawn on a tall page with the text turned 90 degrees.
A quick way to tell the two apart:
- Drag to select a single row of the table. On a truly wide page, the highlight runs left to right along the row. On a rotated page, it may run up or down, or jump between rows.
- Copy a few words and paste them into a plain text editor. Normal words mean the text runs the right way. One letter per line, or words in reverse order, means the text is turned.
- Look at the page size in the document properties. A wide page reports a width larger than its height. A rotated portrait page reports a tall page with a rotation applied.
If the page is rotated, fix the rotation before you convert. Use a rotate command that changes the page itself, not just the view, as described in saving a rotated PDF so the turn sticks. Save a copy, then run the select test again. A scanned landscape page is a different case: it has no text at all, and OCR needs the image upright before it can read any of it.
Wrapped headers in a converted landscape table
Wrapped headers are column labels split across two or three lines because the column is narrow. A converter that reads each line separately produces two header rows, or worse, treats the second line as a first data row. The fix is to merge the lines into one label per column before you build anything on the sheet.
Here is an example. The values are made up. A 12-column landscape report prints its headers in two lines, and the converted sheet shows them as two rows:
| Column | Row 1 in the sheet | Row 2 in the sheet | Header you want |
|---|---|---|---|
| A | Posting | Date | Posting Date |
| B | Value | Date | Value Date |
| C | Description | (empty) | Description |
| D | Ref | No. | Ref No. |
| E | Debit | Amount | Debit Amount |
| F | Credit | Amount | Credit Amount |
| G | FX | Rate | FX Rate |
| H | Orig. | Currency | Orig. Currency |
| I | Fee | (empty) | Fee |
| J | Tax | (empty) | Tax |
| K | Running | Balance | Running Balance |
| L | Branch | Code | Branch Code |
To join them, put this formula in a spare row and fill it across all twelve columns:
=TRIM(A1&" "&A2)
TRIM removes the extra space left where row 2 is empty, as in column C. Copy the results, paste them over row 1 as values only, then delete row 2. You now have one header row of twelve labels, which is what filters, tables and pivot tables expect.
The column-count check
After the merge, run one check. Count the headers in the PDF. Count the non-empty cells in your new header row with =COUNTA(A1:Z1). In the example, both should be 12. If the sheet shows 11, two columns were merged during conversion. If it shows 13, one header was split into two columns, and a data value is probably sitting under the wrong label.
Two-line headers are not the only kind of wrapping. Long descriptions in the body of the table can also wrap, leaving a half-empty row under the real one. Treat those separately, after the headers are fixed, by joining each continuation line to the row above.
How do you handle mixed portrait and landscape pages?
Mixed pages are common in statements and reports: a portrait summary page, landscape transaction pages, then a portrait notes page. Treat each orientation as its own table. Convert the file, find where the column layout changes, and keep the portrait rows out of the landscape table, because their columns do not line up.
The problem with mixed pages is that a converter stacks every page into one sheet. The portrait summary might have three columns: a label, a current figure and a prior figure. The landscape pages might have twelve. Stacked together, the summary values land in the first three columns of the transaction table, under headers such as "Posting Date" and "Value Date" where they make no sense.
A tidy way to separate them:
- Note the page numbers of each orientation while you look through the PDF. For example, page 1 portrait, pages 2 to 7 landscape, page 8 portrait.
- Add a page column if your converter gives you one, or tag rows by hand from the page breaks.
- Move portrait rows to their own sheet. Summary figures belong next to the table, not inside it.
- Check that each landscape page has the same column count. A landscape page that comes out with fewer columns than its neighbours probably had an empty column that the converter skipped.
Some statements are printed in landscape with two narrow pages side by side on one sheet of paper. That looks like mixed orientation but is a different problem, covered in the last section below.
If the conversion fails outright on a mixed file, rather than producing messy rows, the cause may be something else in the file. Checking why a PDF to Excel conversion failed goes through passwords, image pages, damaged files and size limits in order.
Wide landscape sheets in Excel: fitting and checking every column
A wide landscape sheet needs a few Excel settings so you can check it on screen and print it on one page width. Freeze the header row and the first column, narrow the columns that hold short codes, and set the page to landscape with the width fitted to one page. Then compare totals with the PDF.
Viewing first. With twelve or more columns, the running balance on the right is off screen while the date on the left is visible. Select cell B2 and use Freeze Panes on the View tab, so the header row and the date column stay put while you scroll. Zoom out to about 80 percent if the whole row still does not fit.
Checking next. Wide tables give you more places to cross-check:
- Debits and credits should each sum to the totals printed at the foot of the PDF, if it has them.
- Fee and tax columns on a statement often have their own totals. A fee total that is too high may mean a tax figure was read into the fee column.
- The running balance should equal the previous balance plus credits minus debits on every row. A row where it does not is where a value slid into the wrong column.
- Dates in a wide table often come as a posting date and a value date. If they disagree by months rather than days, one of them was read in the wrong day and month order. Fixing converted dates that came out wrong in Excel covers swapped day and month, text dates and lost years.
Printing last. On the Page Layout tab, set Orientation to Landscape. Then open Page Setup and set Fit to 1 page wide, with the height left blank so Excel can use as many pages as it needs. Set row 1 to repeat at the top of each printed page. A twelve-column sheet usually fits on one width in landscape at a readable size.
When a landscape PDF holds two tables side by side
Some landscape pages are not one wide table at all. They hold two narrower tables next to each other, or two portrait pages printed on one sheet. Converting them as one table interleaves the left and right halves row by row. Split the page into halves, convert each half, then stack them in order.
You can spot this layout by its headers. A true wide table has one set of headers across the top. A two-up page repeats the same headers twice: "Date, Description, Amount" on the left and "Date, Description, Amount" again on the right. The page numbers may also give it away, with "Page 3" in the bottom left and "Page 4" in the bottom right.
When this layout goes through a normal conversion, row 1 of the sheet holds the first transaction from the left table and the first transaction from the right table side by side. The dates jump back and forth, and the running balance makes no sense. Sorting by date may seem to fix it, but transactions on the same date can then fall out of their printed order.
The method for this is the same as for any two-column page: crop the left half into one file and the right half into another, convert each, then put the left rows above the right rows. Keeping the halves apart in a two column PDF walks through the cropping, why lines interleave, and how to rejoin the halves in reading order.
FastBankConvert, a bank statement converter, is built to tell you before download if a statement's balances do not add up, and a slid column or an interleaved half is exactly the kind of fault that breaks them. It is getting ready to open; the Get started page is where to join.