When Your Bank Isn't RecognizedWhen a CSV doesn't match any shipped profile, RockSolid Money doesn't guess quietly. It walks you through a short interpretation flow, shows its reasoning, and saves the result so the next file from that institution imports in one step.
1. A proposed mapping
The application reads the header row and the shape of the data and proposes which column is the date, the description, the amount, and so on. You see sample rows from your file interpreted under that proposal — not an abstract settings form.
2. The sign convention is proven, not guessed
The most dangerous thing an importer can do is guess which direction money moved. Where your file carries a running balance column, RockSolid Money checks the arithmetic: if a $50 row makes the balance go down, that row is money out — proven, for every row in the file.
If the file has no balance column and the arithmetic cannot settle it, you're asked in plain language — for a credit card, whether a sample row is a purchase, a payment, or a refund; for a bank account, whether the balance went up or down.
3. Dates are checked the same way
Ambiguous date formats (is 03/04 March 4 or April 3?) are resolved from evidence in the file where possible, and confirmed with you where not.
4. The profile is saved
Once you confirm, the mapping is stored as an import profile for that institution. Future files import against it directly. Profiles are plain data — they describe columns and formats, and can never run code.
If the file still can't be interpreted
It is quarantined with an explanation of what was missing. Nothing partial reaches your ledger. Try a different export format from your bank (OFX/QFX are the most reliable), or email the header row (no amounts needed) to support@rockriverresearch.com and we'll help.
Still need help?
If you're still stuck, contact us or email support@rockriverresearch.com — you will reach the people who build RockSolid Money.