Bank statements

Bank CSV import failed? Column and date fixes for US budgeters

A budgeting app rejected your bank CSV? Check the header row, the MM/DD/YYYY dates, amount signs and stray summary lines, with dollar examples to follow.

Editorial illustration for Bank CSV import failed? Column and date fixes for US budgeters

Nearly every rejected bank CSV traces back to the header row, the date column, the amount column, or extra lines the bank wrapped around the data. Open the file in a spreadsheet, look at the first ten rows and the last five, and you will usually spot the culprit in under a minute.

What does the app actually want from a CSV?

An importer is a very literal reader. It expects one header row, then one transaction per row, with the same number of columns all the way down. It looks for a date, a description and either one amount column or a pair of debit and credit columns. Anything that breaks that pattern gets the whole file refused, even if 99 rows out of 100 are fine.

Say your checking export has 120 transactions. If the bank adds a line at the top reading Account summary, a blank row, and a closing balance line at the bottom, the importer sees 123 rows and cannot decide which one is the header. Deleting those three lines is often the entire fix.

Fix 1: the header and the stray lines

  1. Open the CSV in Excel or Google Sheets, not in a word processor, so you can see rows and columns.
  2. Delete anything above the real header row: account names, date ranges, opening balances.
  3. Delete anything below the last transaction: totals, closing balances, disclaimers.
  4. Make sure the header cells have no blanks. A blank header over a column makes some importers stop.
  5. Save as CSV again, keeping the same file type.

Fix 2: why does a date like 04/05/2026 break the import?

In US exports the order is usually month, day, year, so 04/05/2026 means April 5. An importer set to day-first reads it as 4 May. A date such as 04/25/2026 then fails outright because there is no month 25. If your file has any day above 12 in the first position, you know the order for certain.

The cleanest repair is to convert the column to an unambiguous format. In a spreadsheet, select the dates, then format them as YYYY-MM-DD. If the cells are text rather than true dates, a helper column can rebuild them. For text like 04/25/2026 in cell A2, this works in Excel and Google Sheets:

  • =DATE(RIGHT(A2,4), LEFT(A2,2), MID(A2,4,2)) builds a real date from month, day and year pieces.
  • Format the result as YYYY-MM-DD, copy it, and paste as values over the old column.

Fix 3: amounts that look like numbers but are not

Amounts cause the second most common failure. Watch for these shapes in a dollar file:

What you seeWhy it failsRepair
$1,250.00Currency symbol and comma make it textFind and replace the $ and the comma with nothing
(42.10)Brackets mean negative in accounting styleReplace with -42.10
42.10 CRA suffix instead of a signMove CR or DR into a sign: credits positive, debits negative
1250.0 and 1,250.00 mixedTwo formats in one columnSelect the column, set format to Number with two decimals, and recheck

If the export gives separate Debit and Credit columns, many importers want one signed amount instead. A helper column does it: =IF(D2<>"", -D2, E2) where D holds debits and E holds credits. Spend becomes negative, money in stays positive. Then paste the result as values.

Does the file encoding or delimiter matter?

Sometimes. If merchant names show odd symbols such as a question mark in a diamond, the file is probably not UTF-8. Re-save from your spreadsheet as CSV UTF-8. If everything lands in one column when you open it, the separator may be a semicolon or tab instead of a comma, and the importer may assume the wrong one.

What if only some rows fail?

Then the file is structurally fine and individual rows are the problem. Sort by amount and look for blanks, then sort by date and look for rows that are not dates, such as a note a bank inserted mid-file. Remove or fix those rows, and keep a copy of the untouched original so you can compare later.

SpendixAI reads Excel and CSV exports, including merchant, type, status and currency columns, and lets you review rows before saving, which makes a stray row easy to catch. It is built around India and INR, so mention of it here is only for readers who also keep Indian accounts; no layout is promised to parse perfectly, and the same cleaning habits apply.

A five-minute checklist before every import

  • One header row, no blanks in it.
  • No summary or disclaimer lines above or below the data.
  • Dates in one consistent order, ideally YYYY-MM-DD.
  • Amounts as plain signed numbers, no $, commas or brackets.
  • Row count in the file matches the count you can see on the bank's screen for that period.
  • Original file saved untouched in a separate folder.
SPENDIXAI JOURNAL

Useful ideas start with a clearer picture of your own spending.

Explore your reports