Can I use an Excel file instead of CSV?

Export to CSV UTF-8 first. Spreadsheet files carry formatting, formulas and multiple sheets that make the underlying values ambiguous, and Excel's own date interpretation is a common source of shifted day and month values.

My bank export has one column per row. What went wrong?

The separator was not detected, usually because the file uses a separator that is not a comma, semicolon or tab, or because the export is fixed-width text. Re-export choosing CSV and one of the supported separators.

Does SettlePair upload my CSV files?

No. Files are read with the browser's File API and parsed in the tab; nothing is transmitted or stored, and closing the tab clears the session.

Data preparation

How to prepare CSV files for reconciliation

SettlePair Editorial · Published August 29, 2026 · 3 min read

A reconciliation is only as good as the two files behind it. This is a checklist for exporting invoice and bank data that matches cleanly, and for spotting the format problems that silently distort results.

Key takeaways

You need five fields from invoices and three from the bank; anything else is optional context.
Ambiguous numeric dates are the most common silent error — decide the format explicitly rather than letting a tool guess.
Export as UTF-8 CSV, keep one header row, and quote any value containing the separator.
Check accepted and rejected row counts before trusting any match output.

The columns you actually need

Bank export: transaction date, description or reference, and amount, with deposits positive. Balance columns, categories and running totals are harmless but unnecessary.

Invoice export:

Invoice number — unique, exactly as it appears on the document the customer received.
Customer name — the ledger name; note the trading name they pay from if it differs.
Invoice amount — the gross amount due, as a plain number.
Issue date.
Due date.

Formats that cause silent errors

Dates: 03/04/2026 is ambiguous. Prefer ISO (2026-03-04), or state whether the file is MM/DD or DD/MM before importing.
Amounts: keep them numeric. Currency symbols, thousands separators and trailing footnote markers all have to be stripped, and brackets usually mean negative.
Negative values: (85.00) and -85.00 mean the same thing; be consistent within a file.
Encoding: export UTF-8. A byte-order mark is tolerable, but mixed encodings corrupt customer names with accents.
Separators: comma, semicolon and tab are all workable. A pipe-separated or fixed-width export is not a CSV.

Structural rules

Exactly one header row, with a unique name for every column.
No report titles, subtotal rows, page footers or blank separator blocks.
Any value containing the separator, a quote or a line break must be wrapped in double quotes, with literal quotes doubled.
One record per row — no merged cells or continuation lines carried over from a spreadsheet layout.

A safe row

Invoice No,Client,Amount,Issued,Due
INV-1042,"Harborline Studio, Ltd",1240.00,2026-03-04,2026-04-03

Check before you match

Compare the row count in the export to the record count in the source system, and the file total to the ageing total. If either differs, the export is filtered or truncated and the reconciliation will be wrong in a way matching cannot reveal.

Then read the import summary: accepted rows, rejected rows and the reason for each, duplicates found, and the total value read. Rejected rows must be explained, never ignored.

What SettlePair checks on import

SettlePair validates the file before matching: extension and size, binary content, missing or duplicate headers, a single column caused by the wrong separator, unterminated quotes, rows whose width does not match the header, empty rows, and the row cap. It reports the separator it detected and lists each problem with the row number and a suggested fix.

Row-level problems — unreadable amounts, invalid dates, missing invoice numbers, exact duplicates — are listed individually with accepted and rejected counts. Nothing is discarded silently, and every file is parsed in your browser tab.

Common questions

Can I use an Excel file instead of CSV?: Export to CSV UTF-8 first. Spreadsheet files carry formatting, formulas and multiple sheets that make the underlying values ambiguous, and Excel's own date interpretation is a common source of shifted day and month values.
My bank export has one column per row. What went wrong?: The separator was not detected, usually because the file uses a separator that is not a comma, semicolon or tab, or because the export is fixed-width text. Re-export choosing CSV and one of the supported separators.
Does SettlePair upload my CSV files?: No. Files are read with the browser's File API and parsed in the tab; nothing is transmitted or stored, and closing the tab clears the session.

Sources

General guidance only, not accounting, tax or legal advice. Confirm treatment with your accountant and your own policies.