Roifile

Error codes · EBA technical checks (reception)

306: XBRL instance documents must use UTF-8 encoding.

Rejects the file also shown as EBA.306

Official text

XBRL instance documents must use UTF-8 encoding.

How it appears in feedback

What triggers it

A CSV or JSON file in the package is not UTF-8 (typically ANSI/Windows-1252 saved by Excel). Since 1 May 2025 this is a blocking reception check, because non-UTF-8 text was corrupting integrated data. A non-UTF-8 file also tends to cascade into 807/809 errors.

A file in the package is not valid UTF-8. Typically a CSV saved by Excel as “CSV (comma delimited)” on Windows, which uses the local code page, so names with é, ö or ß break.

Templates: all table files, report.json, parameters.csv, FilingIndicators.csv

How to fix it

Save every file as UTF-8. The EBA slides say UTF-8-BOM is also accepted at the ESAs, but the Central Bank of Ireland guide reports b_01.03.csv in UTF-8-BOM 'will cause errors' on its portal and tells filers to check 'UTF-8 and not UTF-8-BOM'; emit UTF-8 without BOM to be safe everywhere.

Save as “CSV UTF-8 (comma delimited)” in Excel, or convert with a text editor. A UTF-8 byte-order mark is accepted by the ESAs. Since 1 May 2025 this check rejects the whole file.

Conflict recorded verbatim: EBA observations slide 30: 'We will also accept UTF-8-BOM files as this is mandated by EBA filing rules'; CBI guide rule 306 example: 'The table file b_01.03.csv has been encoded in UTF-8-BOM format and this will cause errors in the validation of the file'.

Sources

Check a file for this