Some bank downloads have no header row at all, and that alone can make a budgeting app reject the file. Others carry two or three lines of account details above the data. Either way, an importer cannot tell what each column means, so it stops. Add or tidy the header, and the same file often imports first time.
Does your file even have a header row?
Open the file and read row 1. If it holds a date, a description and a number, then it is data, not a header. Some exports start straight at the first transaction. Insert a row above and type column names such as Date, Description, Amount. Keep the names short and plain, and do not leave any of them blank.
| Row 1 in your file | What it tells you | What to do |
|---|---|---|
| Date, Description, Debit, Credit, Balance | A proper header | Leave it, check the rest |
| 03/11/2026, Grocery, 52.18, , 1,204.33 | First transaction, no header | Insert a header row |
| Account: Chequing, Period: October | A description line, not a header | Delete it and the blank line after |
What if debit and credit sit in separate columns?
This is common, and many importers want a single signed amount. Take a hypothetical month: a C$3,200 pay deposit, C$1,450 rent and a C$63.75 grocery run. In separate columns the pay sits under Credit and the rent and groceries under Debit. A single Amount column should read 3200, -1450 and -63.75.
- Add a column called Amount next to Credit.
- If Debit is column C and Credit is column D, enter =D2-C2 and fill down.
- Check the three lines above produce 3200, -1450 and -63.75.
- Copy the column and paste as values, then remove Debit and Credit.
- Keep the Balance column only if the importer asks for it; some get confused by it.
Why does the date order matter so much?
An export may use YYYY-MM-DD, DD/MM/YYYY or MM/DD/YYYY depending on how the account is set up, and you cannot tell by eye from a date like 03/11/2026. Look for a day above 12 anywhere in the column. If you see 27/10/2026, the order is day first; if you see 10/27/2026, it is month first.
Once you know, convert everything to YYYY-MM-DD. This format sorts correctly, reads the same everywhere and removes the guesswork. Excel and Google Sheets both let you set a custom format of that shape on a column of true dates.
Are there hidden problems in the amounts?
Look for these: a dollar sign, a thousands comma, brackets for negatives, a trailing minus, and a text note like Pending in the amount cell. If you sort the amount column and the text rows jump to one end, you have found them. Fix with find and replace, and decide how to treat pending lines: leaving them out and importing again after they post avoids duplicates.
Test small, then the full file
Copy the header and the first ten rows into a new file and import that. If it works, import in larger pieces. If it fails, the problem is in those ten rows, which is far easier to see than in 400. Keep the full original download untouched so you can start again.
SpendixAI reads Excel and CSV exports including merchant, type, status and currency columns, and shows rows for review before saving. It is built around India and INR rather than Canadian accounts, so consider it only if you also track Indian ones; no file layout is promised to parse, and the cleaning steps here help with any importer.
Pre-import checklist
- A single header row with no blank names.
- Nothing above the header and nothing below the last transaction.
- One signed amount, or debit and credit columns the importer explicitly supports.
- Dates in YYYY-MM-DD.
- Row count and total match what the bank screen shows.
Useful ideas start with a clearer picture of your own spending.
Explore your reports

