If a transaction description on your Avanza statement is long enough to wrap across two or three lines, it still needs to land in a single row once converted. When a transaction description spans several lines, the raw table can splinter a single entry across two or three rows. Convert·Into reads the full block and stitches those wrapped lines back together, so each transaction lands on one row with its date, payee and amount intact. The engine does not rely on row counts or fixed cell heights; it follows the text flow and reassembles the record before mapping it to the export columns. A split entry in a statement document comes out as one clean transaction in the output, ready for the ledger.
When your Avanza statement arrives as a scan or a photo of the page rather than a clean export, the numbers on it are just pixels until they are extracted. If the statement was scanned from paper or captured with a phone camera, the image quality is rarely perfect. Convert·Into applies OCR to read every character on the page, then checks each figure against the statement's own arithmetic. A smudged digit, a tilted scan or a shadow across the print rarely produces a clean read on the first pass, but the running-balance verification catches the mismatch and flags the row. You review the flagged line, correct the misread number, and the rest of the file stays intact.
Whatever the running balance looks like on your Avanza statement, Convert·Into checks it against the statement's own printed closing figure before you export anything. Every statement carries a printed closing balance. Convert·Into recalculates the running balance line by line from the extracted transactions and compares the result against that closing figure. Anything that does not add up is flagged for review instead of rejecting the whole file, so a single suspect row never stops the export. You confirm or correct the flagged entry and move on; the rest of the data is already verified.