Converting COREP to XBRL-CSV
The conversion path the validation report sits inside.
Read the tutorial →Every conversion produces a full validation report, whether or not the run is clean. This explains what each layer tests, how to read a finding, and the order that clears the largest number of errors for the least work.
Validation is not a single test. Findings are produced by six layers applied in sequence, and knowing which layer produced a finding usually tells you where the fix belongs.
| Layer | What it tests | Typical finding |
|---|---|---|
| Structure | That the input parses and carries the sheets, columns or elements the framework expects. | A missing table, an unreadable file, a column that never arrived. |
| Data type | That each value is of the declared type, length and precision. | Text in a numeric field, a date in a local format, excess decimal places. |
| Controlled vocabulary | That coded fields carry values from the applicable list. | A country name instead of a code, a lapsed identifier, an unrecognised type code. |
| Referential integrity | That every reference between records resolves. | A row pointing at a parent record that is not in the dataset. |
| Framework rules | The arithmetic and logical rules published for the framework — totals, cross-template consistency, permitted combinations. | A subtotal that does not sum, a combination the framework forbids. |
| Cross-period | Consistency with what you filed for earlier periods, where the framework compares them. | An opening figure that does not match the previous closing figure. |
Each finding carries the same elements, and reading them in order is faster than reading the message first.
Findings that arrive in clusters usually share one cause. A single unresolved reference or one mistyped identifier will produce a finding for every record that depended on it, so counting distinct causes matters more than counting findings.
A blocking finding means the framework’s own rules are not satisfied and the output package is not produced. A warning means a rule flagged the value as improbable rather than impermissible — a large period-on-period movement, an unusual but permitted combination, an empty section that is optional.
Warnings should not be dismissed as noise. Receiving systems run their own plausibility checks, and a warning here is frequently the same check the authority will raise as a query later. The useful discipline is to resolve each warning to one of two states: corrected, or explained and recorded.
Working through a report in the order findings are listed is the slowest possible route. This order clears the most for the least effort:
Every finding resolves in one of three places, and misplacing the fix is what makes a filing cycle repeat.
| Fix location | When | Consequence of fixing elsewhere |
|---|---|---|
| Source data | The value is wrong or missing in the system of record. | Correcting it in the upload means the same finding returns every period. |
| Mapping | The value is correct but is expressed in your own vocabulary rather than the framework’s. | Editing individual cells hides a mapping gap that will grow. |
| Judgement | The rule is satisfied but the figure is unusual, or a classification is genuinely arguable. | Silently adjusting data to clear a warning removes the record of the decision. |
The validation report is part of the filing record. It shows what was checked, what was found and what state the data was in when the package was produced — which is exactly what an authority, an auditor or a successor team asks for when a figure is questioned much later. Download and retain the report for every submitted run, including the runs that were superseded, and keep the explanation attached to any warning that was accepted rather than corrected.
Primary instruments only. Each is named in full so the reference remains traceable even if a link moves.
REGREP is an independent software provider. This record explains a reporting framework in plain language and is not legal, tax or regulatory advice. Confirm scope, thresholds and submission dates with your competent authority before you file.
More on this framework, and the module that produces the filing.
The conversion path the validation report sits inside.
Read the tutorial →Frameworks covered, checks applied and output formats produced.
See capabilities →How uploaded data is stored, separated and retained.
Read about security →The free tier produces the complete validation report on your own data, with every finding traced to the cell it came from.