> ## Documentation Index
> Fetch the complete documentation index at: https://docs.budgetbandit.io/llms.txt
> Use this file to discover all available pages before exploring further.

# CSV format

> Reference for the columns, date formats, and amount handling Budget Bandit expects when importing a CSV.

This is the reference for what a CSV needs to import cleanly. For the step-by-step
flow, see <a href="/import/csv">Import a CSV</a>.

<Note>
  You do not have to rename your bank's columns. The importer lets you map any
  column to the right field, and auto-detects the common ones. This page tells
  you what each field expects.
</Note>

## Required fields

A CSV must provide:

* A **Date** column, and
* Either an **Amount** column, **or** an **Inflow** and **Outflow** pair.

Without both, the import cannot proceed.

## Fields you can map

| Field                   | Required                        | Notes                                                     |
| ----------------------- | ------------------------------- | --------------------------------------------------------- |
| **Date**                | Yes                             | One date per transaction.                                 |
| **Amount**              | One of Amount or Inflow/Outflow | Single signed column. Negative is spending.               |
| **Inflow (credit)**     | —                               | Money in. Use with Outflow when your bank splits the two. |
| **Outflow (debit)**     | —                               | Money out.                                                |
| **Payee / Description** | No                              | Merchant or description.                                  |
| **Memo / Notes**        | No                              | Extra text.                                               |
| **Category**            | No                              | A category name if your file has one.                     |
| **Reference ID**        | No                              | A bank-supplied unique ID; improves duplicate matching.   |
| **Skip**                | —                               | Mark columns you do not want imported.                    |

You cannot map both a single **Amount** column and an **Inflow**/**Outflow**
pair — choose one approach. The importer disables the conflicting options for
you.

## Date formats

Leave the date format on **Auto-detect** unless the sample looks wrong. The
supported explicit formats are:

* `MM/DD/YYYY`
* `DD/MM/YYYY`
* `YYYY-MM-DD`
* `M/D/YY`
* `M/D/YYYY`
* `DD-MM-YYYY`

If your dates are ambiguous (for example `03/04/2026`), pick the exact format so
day and month are not swapped.

## Amounts

* A single **Amount** column should be **negative for spending** and positive for
  money in.
* If your bank reports spending as positive numbers, tick **Invert amounts** to
  flip every sign.
* If you map **Inflow** only, every row imports positive. **Outflow** only,
  every row imports negative.
* Currency symbols and thousands separators (like `$1,250.00`) are handled.

## Duplicate matching

A **Reference ID** column, where your bank provides one, gives the most reliable
duplicate detection — exact ID matches are caught precisely. Without it, Budget
Bandit falls back to matching on date, amount, and payee within a few days. See
duplicate handling in <a href="/import/csv">Import a CSV</a>.

## A minimal example

```csv theme={null}
Date,Description,Amount
2026-06-01,Coffee Shop,-4.75
2026-06-02,Paycheck,2200.00
2026-06-03,Grocery Store,-86.20
```

That file imports cleanly: map `Date` to Date, `Description` to Payee /
Description, and `Amount` to Amount.

## Related

<CardGroup cols={2}>
  <Card title="Import a CSV" icon="table" href="/import/csv">
    The full import flow.
  </Card>

  <Card title="Why rows did not import" icon="circle-question" href="/import/why-rows-didnt-import">
    Diagnose skipped or dropped rows.
  </Card>
</CardGroup>