Whichever order your Julius Baer Group statement prints its columns in, each one still has to land in the right place once it reaches a spreadsheet. The column layout is identified during extraction. A transaction table drawn on a statement page can present columns in almost any order: date, description, debit, credit, balance, or sometimes value date before transaction date, or a combined amount column that uses a sign to indicate direction. The engine maps each detected column into a standard field so the data arrives in a consistent structure. Debits and credits are separated into their own columns; date formats are normalised; and any column that the statement itself labels with a non-English heading is translated to its English ledger equivalent. If the extraction places a column in the wrong slot, you can adjust the mapping before export. The chosen mapping is saved, so the next statement document that shares the same layout follows the same assignment without further adjustment.
If your Julius Baer Group statement runs to several pages, uploading it starts the same layout detection Convert·Into applies to any statement document, page by page. The engine handles statements that run several pages long. When the transaction table breaks across pages, interrupted by page headers, footers, carried-forward subtotals or blank space, the rows that resume on the next page are joined to the same table without duplication. No transaction is split, and no subtotal row is mistaken for a new entry. The output is a single flat table, one row per transaction, regardless of how many pages the original document spanned.
Whatever the running balance looks like on your Julius Baer Group statement, Convert·Into checks it against the statement's own printed closing figure before you export anything. Every statement is checked against its own printed closing figure. The running balance is recomputed row by row from the extracted amounts, and the final computed balance is compared to the closing balance shown on the original document. If a row does not add up, it is flagged rather than blocking the entire file, so you can review the single problematic line while the rest of the data is ready for export. Anything that does not match is held for your inspection, not silently passed through.