Import Problems
Messages about the whole file
“PDF statements cannot be imported — download a CSV, OFX, or Excel export from your bank instead.”
A PDF is a picture of a table. Look in your bank for “Export”, “Download”, or “Activity” rather than “Statements” — most banks offer CSV somewhere even when they lead with PDF.
“Legacy .xls workbooks are not supported — re-save the file as .xlsx or CSV.”
Open it in Excel, LibreOffice, or Google Sheets and save it again as .xlsx or CSV.
“Unrecognised binary file.”
The file is not a format Aventurine reads. The supported list is in Supported File Formats.
“The file appears to be empty or has no data rows.”
The file opened but had nothing usable in it. Common causes: the export produced a header row and nothing else, or the date range you asked your bank for had no transactions.
“Could not read the file.”
The file could not be opened at all — it may have been moved, deleted, or be zero bytes. Try downloading it again.
“The file is not valid JSON” / “No array of transaction objects found in the JSON.”
For JSON imports, Aventurine expects either a list of transaction records or an object containing one. If you exported from another app, check that you got the transactions file rather than a settings or metadata file.
“No worksheet found — is this really an .xlsx workbook?”
The file has the shape of an Excel workbook but no readable sheet in it. Re-save it from a spreadsheet program.
Messages about individual rows
These appear as a banner on the Here Is Your Import screen, naming a row number and a reason. Those rows are skipped; everything else imports normally.
“unparseable date”
The date column contains something Aventurine cannot read as a date. Usually this means the wrong column is mapped to Date — check the preview on the Map Columns screen. It can also be a summary or total row at the bottom of the file, in which case skipping it is correct.
“unparseable amount”
Same idea for the amount column. Also common on total rows and on rows where the bank wrote something like “PENDING” in place of a figure.
“empty description”
Every transaction needs a description. Either the wrong column is mapped, or that row genuinely has none — often a blank separator row.
“empty amount”
In a Debit/Credit file, neither column had a value.
“both debit and credit have values”
A row has a real number in both the money-out and money-in columns, so the direction would be a guess. Aventurine refuses rather than pick one. Check whether the two columns are mapped correctly; if the file really contains such a row, that row will need adding by hand.
“No rows could be read”
Every row failed, so the import stops with the first error shown. This nearly always means the columns are matched wrongly — go back and check the preview.
Things that are not errors
“23 transactions will be left uncategorized.” Expected. Aventurine only fills in categories it is sure of. See Automatic Categorization.
Rows badged “duplicate”. They already exist in your ledger and are unticked so they will not be added twice. See Duplicate Transactions.
After the import
“Import failed: …” appears if the save itself did not work. The review screen comes back so you can try again. Nothing partial is written — either the whole import lands or none of it does.
Problems that appear later
My transactions imported but the dashboard is empty.
The dashboard opens on the current month. If you imported older history, step the month picker back. See The Dashboard.
My transactions are in the wrong account.
There is no bulk “move to another account”. Filter to the affected rows on the Transactions page, select them, and change the account in the edit dialog. If there are many, deleting them and re-importing is faster.
Everything came in as an expense (or as income).
The amount column’s signs were not read as you expected. If your file uses separate Debit and Credit columns, check they were both mapped — see Matching Your Columns. Delete the rows and import again with the correct matching.
Amounts are 100× too big or too small.
A European-style file (1.234,56) read as US style, or the reverse. Aventurine works this out across the whole file, so it is rare — but if you see it, check whether the file mixes conventions. Delete and re-import after saving the file in a consistent format.
The same transaction appears twice with slightly different text.
Your bank changed the description between the pending and settled versions, so duplicate detection did not recognize them as the same. Delete the extra. Importing only settled transactions avoids it.
Related: Importing Transactions · Supported File Formats · Troubleshooting