DATA AND BACKUPS
Clean a transaction CSV and protect the original
Map columns, check timestamps and duplicates, and keep a recoverable copy before importing transactions.
Open the related tool →Keep the original export untouched
Before cleaning a transaction file, keep an unedited copy from its source. Work with a second file so you can investigate any mapping mistake later. A filename that includes the account label and export date is more useful than repeatedly overwriting transactions.csv. Do not include secrets or exchange API keys in the file.
Understand the columns before mapping them
The importer distinguishes date, asset, type, quantity, price, fee, external ID and notes. Price is a per-unit figure in USD, not necessarily the transaction’s total value. If your source contains only a total, review how it should be converted before importing. For two units at $100 each, the trade value is $200, not a $200 unit price.
| Source field | Interpretation to verify |
|---|---|
| Timestamp | Includes an offset, or needs an explicit timezone |
| Symbol | Maps consistently to the intended asset ID |
| Quantity | Number of units, not cash value |
| Unit price | USD per unit |
| Fee | USD cost under the source export’s definition |
| External ID | Stable identifier for the original record |
Preserve valid CSV quoting
A note containing a comma must remain one field. In common CSV syntax, wrap that field in double quotes and represent an embedded quote by doubling it. A quoted field may contain a line break. Splitting every physical line or comma manually can therefore damage a valid file.
date,asset,type,quantity,price,fee,externalId,notes 2026-10-07T09:00:00Z,bitcoin,buy,2,100,1,example-001,"Example, not a real trade"
Preview before merging
- Select the CSV and check the column mapping.
- Choose the timezone for timestamps that have no offset.
- Review asset aliases and rejected rows.
- Inspect accepted quantities, prices and fees against the source.
- Merge once, then inspect the saved transactions.
Recognized duplicates are skipped, but a changed timestamp, amount or identifier can describe a different record. Duplicate detection is a safeguard, not a substitute for reviewing the original export. The transaction list does not automatically update holdings or diary entries.
Make a restorable backup
Download a full JSON workspace backup before a large import or replacement restore, and confirm that it exists outside the browser. Browser storage can be cleared or evicted. CSV and Excel reports are useful for analysis, but they do not contain every workspace setting and record family needed for a full restore.
Frequently asked questions
Can I reuse my column mapping?
Save an import profile and select it manually for a later file with matching headings.
Will importing twice always add no rows?
Recognized equivalents are skipped. Changed source fields can be distinct, so inspect the preview and merge counts.
Sources and further reading
Sources checked 7 October 2026. Worked examples are independently calculated illustrations.
Educational examples only. No trades are executed. Check source data and assumptions before relying on a result.
Try this in the tool →