How to total a bank export's decimal-comma amounts
European bank exports write -1.234,50, -12 480,00: a comma for the decimal, a dot or a space for the thousands. Here's why that breaks a sum in most tools, and how Tabulay reads the column as what it is, a number, sorts it, totals it, and saves it back exactly as written.
Why a bank export's amounts don't add up
A European bank or payment export writes its amounts the way the country does: a comma for the decimal mark, a dot or a space to group the thousands, so a transaction reads -1.234,50 or -12 480,00. Most spreadsheets, opened with their own regional number format, expect the opposite convention. The column still looks like a list of numbers; it just isn't one to the software reading it. A SUM comes back wrong or blank, a header sort puts -12 480,00 above -1.234,50 because it's comparing text, and a filter for "over 1,000" misses every row that matches. Finance and ops teams end up checking the export against the paper statement by hand, line by line.
Detected on open
Open the file (File ▸ Open with Options, ⇧⌘O) and Tabulay detects the decimal separator per column, along with the delimiter, encoding and date format, with a live preview before a single row opens. A column written -1.234,50 or -12 480,00, thousands grouped by a dot or a space, becomes a real number column: dec(12,2) in its two-line header, "Decimal comma" in the inspector.
Sort, filter and total it, free
- Sort the header, and -12 480,00 sorts below -1.234,50, numerically, not alphabetically.
- Filter chips and a right-click filter ("over this amount") read the real value, free for everyone.
- Select the column, and the status bar shows its sum, average and count in exact decimals, decimal commas included, up to 200,000 rows of the view (it says so when it stops there), with a warning if currencies are mixed.
Total the whole file, or fix a stray value
For the whole file, not just the rows in view, SQL (Tabulay Pro, included in the 14-day trial) totals every row: SELECT sum(amount) FROM this reads the decimal-comma column as a number, because it's typed from the file's own schema, the same way amount < -1000 would be. When a handful of rows in a decimal-point column were typed the European way by mistake, Normalize (Tabulay Pro, included in the 14-day trial) rewrites only those values, -1.234,50 to -1234.50, and shows the exact count first.
Saved exactly as written
Saving doesn't touch the notation. Untouched rows are copied byte for byte, and an edited row is written back with the file's own delimiter and encoding: -1.234,50 stays -1.234,50 on disk, unless you change that cell yourself.
The limits, stated plainly
A column reads as a number only when its notation is consistent. Amounts with a thousands comma (1,234.50), a currency symbol (€1.234,50), or a trailing minus (1234,50-) stay text: filter and sort them as text, or normalize them into a number column first. A column whose values could mean either notation (1.234, or 12) keeps the decimal point rather than being guessed at. The status bar totals up to 200,000 rows of the view; past that, only SQL totals the rest. Tabulay doesn't touch a value's own notation or currency: it reads exactly what's written, totals it, and saves it back unchanged.
Other ways to total a bank export
A command-line tool can total a decimal-comma column too, once the notation is converted by hand: swap the comma and the dot, strip the thousands separator, then sum the result. That's a script to write and test for each file's dialect, with no grid to check a total against the original rows.
A spreadsheet can read the column correctly if its own number format, or the system's regional settings, match the file before it's opened. That works well enough for a file you already trust the shape of; for a file with mixed formats, for millions of rows, or for the rows you'd rather not touch by hand, it's slower, and easier to get quietly wrong.
Try it
Open a bank export with File ▸ Open with Options and watch the decimal comma get detected before a single row shows. The full list of number formats is on the tech specs page, and running SQL on a CSV file covers totaling the whole file in more depth.