Open the statement file, find the row that holds the real column names (Date, Narration and so on), delete everything above it, and make sure only that one row acts as a header. Do it on a copy, and the same file loads cleanly into a table, a pivot or Power Query afterwards.
Why do statement files start with junk rows?
A statement is designed to be printed, so the top of the file usually carries your name, an account number, the period covered and a few blank rows before the table begins. A spreadsheet needs the opposite: one header row, then data, nothing else. Left as it is, sorting, filtering and pivots treat the account details as if they were data.
Before you delete the account-details rows, read them once. The period covered and the opening balance are useful later, when you check that the months join up. Copy those two lines to a notes sheet, or type them beside the file name, and then delete. It costs thirty seconds and saves hunting through the original download months afterwards.
How do you delete the rows by hand?
- Save a copy of the file first, with a name such as statement-raw.
- Click the row number of the first row above the header, hold Shift and click the row number of the row just above your header row. Everything above the header is now selected.
- Right-click any selected row number and choose Delete. The header moves to row 1.
- Scroll through the first twenty rows and confirm that no second header or blank row remains.
Worked example with invented numbers: say the file has 9 rows of account details and blanks, the header on row 10, and the first transaction on row 11. Select rows 1 to 9, delete, and the header becomes row 1. If the file had 240 rows in total, you should now see 231, because 240 minus 9 is 231. Checking that count is a quick way to make sure you deleted what you meant to.
Can Power Query skip the top rows for you?
Yes, and it is the better choice if you get a file every month. Choose Data, Get Data, From File, From Text/CSV (or From Workbook for Excel files), then Transform Data. In the editor, use Home, Remove Rows, Remove Top Rows and type the number of junk rows, then Transform, Use First Row as Headers. The steps are recorded, so next month's file with the same layout takes one refresh.
What if the number of junk rows changes every month?
Then a fixed count is risky. Two ways around it: filter the first column in Excel to find the row whose value is the header word, note its row number, and delete up to that point; or in Power Query use a condition instead of a count. Replace the fixed number in the Remove Top Rows step with Table.Skip(Source, each [Column1] <> "Date"), which skips rows until the first column reads Date. Check the wording against your own file first, because your bank may print the first heading differently.
What do you check after cleaning?
- Row 1 holds the column names and row 2 holds the first transaction.
- Every heading is unique and non-empty, which a Table needs.
- The date column is formatted as dates, not left-aligned text.
- The row count matches your before-and-after arithmetic.
- Footer lines at the bottom are handled separately, since they sit after the data, not above it.
A closing sentence for the app side of things: SpendixAI takes a PDF, CSV or Excel statement directly and shows the rows for review before saving, so for one-off imports the header clean-up is not something you need to do yourself, though a layout it cannot read may still need a look.
Useful ideas start with a clearer picture of your own spending.
Explore your reports


