Data prep
How to Clean a Bank CSV Before Uploading
A few quick fixes can prevent upload errors and improve recurring detection. This walkthrough keeps things simple so your file is ready for the detector, at a pace that feels comfortable.
Published: December 24, 2025 - Updated: December 24, 2025 - By FinToolSuite Editorial
Quick answer
- Keep one header row.
- Make dates consistent.
- Ensure amounts are numbers (no currency symbols).
- Remove duplicates and empty rows.
- Remove sensitive columns.
Disclaimer
Educational purposes only; not financial advice. Examples are illustrative and results depend on your data; they are not guaranteed. Do not upload sensitive data you do not need (account numbers, addresses, reference IDs). Review your file before uploading.
Your data is processed to detect recurring patterns; see the Privacy Policy for details.
Start with a copy (safety tip)
Save a copy of the original export. Clean the copy, not the original. A spreadsheet app or text editor works fine.
Step by step cleaning checklist
- Keep only the needed columns.
- Date
- Description
- Amount
- Category and Balance are optional.
- Remove extra header rows and empty rows.
Banks sometimes include top notes or blank lines; delete them so the header starts at row 1.
- Unify date format.
Choose one format and apply it to the whole column; avoid mixing 24-12-2025 and 12-24-2025.
Before: 24-12-2025, 12/25/25
After: 2025-12-24, 2025-12-25 - Fix amount formatting.
- Remove currency symbols.
- Remove thousands separators if needed.
- Convert commas as decimals to dots if required.
- Keep negatives for spending if your export uses that. Sign rules: see the CSV format guide.
- Handle pending and duplicate transactions.
Pending lines can appear, then appear again as posted. Choose a rule: keep posted, remove pending. If you see exact duplicates, remove one.
Issue What you see Simple fix Pending + posted Same amount twice Keep posted, remove pending. Exact duplicate Identical row repeated Remove the duplicate row. - Fix separators and encoding if upload fails.
- Check commas vs semicolons.
- Save as CSV UTF-8.
- Ensure one consistent delimiter.
Mini sample rows
Date,Description,Amount
2025-11-01,STREAMING SERVICE,-9.99
2025-11-03,GYM MEMBERSHIP,-25.00
2025-11-05,COFFEE SHOP,-4.20
2025-11-08,CLOUD STORAGE,-2.99
2025-11-12,SALARY,2500.00Plain text descriptions and numeric amounts are enough. Consistent dates help the detector group recurring items.
Privacy tips
- Remove account number columns.
- Remove full addresses and reference IDs.
- Limit the date range to what you need.
- Store cleaned files securely.
See the Privacy Policy for data handling details.
Common mistakes
- Leaving currency symbols in the amount column.
- Mixed date formats.
- Multiple header rows.
- Uploading a PDF instead of CSV.
- Keeping both pending and posted duplicates.
Troubleshooting
- Compare your columns to the format guide.
- Try exporting again for a shorter date range.
- Open the file in a plain text editor to check the delimiter.
- Review the detector FAQ for common upload questions.
FAQ
Do I need category columns?
No. Date, Description, and Amount are enough.
How do I handle refunds?
Keep refunds with their sign as exported; they help balance the totals.
What about credit card statements?
If exported as CSV with the needed columns, they work the same as bank accounts.
What if my CSV uses semicolons?
Save with the same delimiter, or re-save using commas for consistency.
Why do I see duplicates after export?
Some exports repeat pending and posted lines. Remove one copy before upload.
Are results guaranteed?
No. Flags are illustrative; confirm with your statements.
Final call to action
Clean your CSV, upload it, and review the detected subscriptions with a calm, consistent process.