SEPA file errors: why a bank rejects your file
Last updated: 28 October 2026
When a bank refuses a SEPA file on submission, the cause is usually one of a handful of errors in the XML. This page lists the six we see most often, with the message you get and how to fix it. Upload your file below for a free check of all of them at once.
Check your file for the errors banks reject
Choose fileChecking is free. Repair if needed: €14.95 excl. VAT per file.
Our server checks your file in memory only, never saves it to disk and deletes it after 30 minutes at the latest.
A rejection on submission happens before anything is collected or paid. The bank checks the file as soon as it receives it and refuses the whole file or a batch at once, reporting a reason code in the portal or in the status file pain.002. That is different from a return afterwards: an R-transaction, where a single direct debit is processed first and comes back later, for example because the account has insufficient funds. Those returns have their own codes; see SEPA reason codes.
The six errors below are the ones our tool finds most often in rejected files, and the ones it can largely repair itself. Each page explains the cause, how to fix the file by hand and what the repair does. The check is free; a repair costs €14.95 excl. VAT per file.
| Error | Typical message | Repair by the tool |
|---|---|---|
| XML schema error | This element is not expected | Automatic for version mix-ups, otherwise reported |
| Control sum incorrect | AM10 InvalidControlSum | Automatic: both totals recalculated |
| Invalid characters | xmlParseEntityRef: no name | Automatic in EndToEndId, PmtInfId and MsgId |
| Date in the past | DT01 InvalidDate | New date on request, never chosen silently |
| Invalid amount or IBAN format | '25,00' is not a valid value of the atomic type | Automatic when unambiguous, otherwise reported |
| Field too long | [facet 'maxLength'] … exceeds the allowed maximum length | Shortened automatically, except the mandate reference |
All file errors
XML schema error
The file does not match the ISO 20022 schema: an element is in the wrong place, carries the name from another version or does not exist. Version mix-ups such as <BIC> instead of <BICFI> are repaired automatically; other ordering errors are reported so you can fix the export.
Control sum incorrect
The number of transactions (NbOfTxs) or the control sum (CtrlSum) in the header does not match the payments. The schema accepts the file, so only the bank or a content check finds it. The repair recalculates both totals at file and batch level.
Invalid characters
A reference contains a character outside the SEPA character set, such as an underscore, or a name contains a bare & that breaks the XML. Characters in EndToEndId, PmtInfId and MsgId are replaced automatically; the mandate reference and invalid XML are left to you.
Date in the past
The requested collection or execution date has passed, or is not written as YYYY-MM-DD. The tool converts an unambiguous notation itself and asks you for a new date when the old one has passed, with the earliest possible date as a suggestion.
Invalid amount or IBAN format
Amounts need a decimal point and at most two decimals; IBANs are written without spaces and in capitals. The tool converts an unambiguous notation itself, normalises IBANs and does not round or guess.
Field too long
A name, remittance information or reference is longer than allowed: 70 characters for a name under the EPC rules, 140 for the description and 35 for references. The tool shortens names, descriptions and EndToEndId, but never the mandate reference.
Other reasons for a rejection
A file can also be rejected for reasons outside these six. A common one is a duplicate reference: banks reject a file with a MsgId or EndToEndId they have already received (AM05 Duplication), for example when the same file is uploaded twice. Our tool cannot see what was sent to your bank before. A duplicate EndToEndId within one file is reported as a warning but not changed, because it is your own reference; a duplicate PmtInfId within the file is repaired automatically.
Other topics have their own pages: an outdated file version, the structured address, the creditor identifier and the mandate reference.
- SEPA XML versions – when your bank no longer accepts an older version such as pain.008.001.02.
- Structured address – how to put addresses into separate fields.
- SEPA creditor identifier – the identifier that every direct debit file must carry.
- SEPA mandate reference – rules for the reference of each mandate.
- SEPA reason codes – returns after a collection, such as AM04 or MS03.
Frequently asked questions
What is the difference between a rejection and a return?
A rejection on submission concerns the file or a batch and happens immediately, before any payment. A return (R-transaction) concerns a single payment that has already been processed; see SEPA reason codes.
Where do I find the reason for a rejection?
In your bank portal or in the status file pain.002, where the reason code is in the Rsn/Cd field. Online banking often shows a translation of it.
My file passes an XML validator. Can it still be rejected?
Yes. A wrong control sum, a date in the past or a name over 70 characters all pass the schema check. Our check also looks at these SEPA rules.
Does the repair change my payments?
It changes the notation, names and totals that cause the rejection and lists each change in the change report. The mandate reference is never changed, and amounts are never rounded.
Sources
- European Payments Council: SDD Core Customer-to-PSP Implementation Guidelines 2025 – fields, lengths and the SEPA character set for direct debit files
- European Payments Council: 2025 SDD Core Rulebook, version 1.1 – rules of the SEPA Core scheme, including the earliest collection date
- ISO 20022: External Code Sets – the list of reason codes such as AM10, FF01 and CH16 used in status reports