What a dimension is
A plain XBRL fact has a concept, an entity, a period and a value. That is enough for “total assets”, but not for “premiums written, motor insurance, in Austria”. Dimensions add the rest.
The specification defines a dimension as one of “the different aspects by which a fact MAY be characterised”, and a domain as the set of members a dimension can take. In an instance the values sit inside the context, in the scenario or segment element and nowhere else (section 2.1).
An example: two contexts from a Solvency II return
These contexts come from the EIOPA public sample for the annual return (ARB, taxonomy 2.10.0). Each fact points at a context by its id, so every number in the return inherits that context’s dimension values.
- Facts and contexts
- 25,737 facts across 13,178 contexts; only 1 context has no dimension values
- Dimensions used
- 100: 52 explicit, 48 typed
- Where they sit
- All in xbrli:scenario, none in xbrli:segment
- What a typed value looks like
- An ID inside xbrldi:typedMember, one per row of a list template
The first fact is broken down by one coded member from a fixed list. The second belongs to one row, named by a free ID, and also carries a coded member. Check each against the hypercube for its concept, and a filing that uses a dimension the cube does not allow is stopped on arrival.
An explicit and a typed context (shortened)ShowHide
<xbrli:scenario>
<xbrldi:explicitMember dimension="s2c_dim:LG">s2c_GA:AT</xbrldi:explicitMember>
</xbrli:scenario>
<xbrli:scenario>
<xbrldi:typedMember dimension="s2c_dim:FN">
<s2c_typ:ID>637lblTh7G6V79TTBJ+Ejw==</s2c_typ:ID>
</xbrldi:typedMember>
<xbrldi:explicitMember dimension="s2c_dim:PO">s2c_PU:x60</xbrldi:explicitMember>
</xbrli:scenario>Contexts copied from the EIOPA ARB 2.10.0 public sample served by Line4 Check, with the scenario shown and the entity and period left out. Codes are the taxonomy’s own; their labels are in the EIOPA taxonomy.
Why every supervisory return is built on them
A template such as an own funds table or a list of assets is a grid. In a data point model each cell is a concept plus a set of dimension values, so the taxonomy can describe thousands of cells with a few hundred concepts and a few dozen breakdowns.
The Table Linkbase then lays those cells out as the rows, columns and sheets firms see. In XBRL International’s words it “provides a mechanism for taxonomy authors to define a tabular layout of facts”. Dimensions say what each cell means; the table says where it is drawn.
Closed hypercubes: what keeps a filing tight
Every all arc must say whether its dimensions go in the segment or the scenario, and it may be marked closed. A closed hypercube allows no other dimension values in that element, so a fact with one extra breakdown fails (definition 21).
Supervisory taxonomies usually close their cubes so each cell has exactly one meaning. That is what lets a validation rule find the right number every time, and why a fact with a stray or misspelt dimension is an error, not a warning.
What it means for supervisors
If you design returns, model every breakdown as a dimension with a named domain, use typed dimensions only for open lists such as assets or counterparties, and close your hypercubes. Firms then have one way to report each cell, and you can compare the same cell across firms and periods.
If you receive filings, check every fact against its hypercube before you load it. A fact with a dimension the cube does not allow, a member outside the domain or a default written out is rejected on arrival with the cell it was meant for, so the firm can fix it the same day.
Where AI fits, and where people decide
AI can explain a dimensional fact in plain words, from its codes and the taxonomy’s labels, and can draft the breakdowns for a new template from the policy text.
Which breakdowns a return needs is a supervisory decision. Whether a fact fits its hypercube is a fixed rule, checked the same way every time and kept with each filing.
- 1Is every breakdown in your templates a dimension with a named domain?
- 2Are typed dimensions used only for open lists, such as assets or funds?
- 3Are your hypercubes closed, so each cell has one meaning?
- 4Does your platform check every fact against its hypercube on arrival?
- 5Do firms get the cell and the dimension in the error message?
Where Line4 fits
Line4 is SupTech software for financial regulators. It collects returns in XBRL, Excel and CSV by portal, API or MCP, validates them against your taxonomies and rules at the point of submission (see XBRL software for regulators), and puts AI to work reading them, with supervisors making every decision. If you work with a standard we have not yet run in production, we will say so plainly and show you how we would approach it.
Questions
What are XBRL dimensions?
Extra aspects of an XBRL fact, beyond concept, entity and period, such as country, line of business or asset. They sit in the context’s scenario or segment element and are defined by XBRL International’s Dimensions 1.0, a Recommendation since 18 September 2006 with errata corrections to 25 January 2012.
What is the difference between an explicit and a typed dimension?
An explicit dimension takes a member from a list in the taxonomy, so the value is a code, written as xbrldi:explicitMember. A typed dimension has no list: the value is free content of a set type, such as an ID, written inside xbrldi:typedMember. Supervisors use typed dimensions for row-per-item templates such as lists of assets.
What is a hypercube in XBRL?
A set of dimensions. A taxonomy links a hypercube to a concept with an all arc, saying which dimensions facts for that concept carry, or with a notAll arc to exclude combinations. Each all arc must name segment or scenario as the place the dimensions go.
What does a closed hypercube mean?
If the all arc has closed="true", no dimension values other than the cube’s may appear in that context element. Facts with an extra breakdown are invalid. Supervisory taxonomies usually close their cubes so every cell has exactly one meaning.
What is a dimension default?
The member assumed when a dimension is left out of a context, set with a dimension-default arc. The default member must not be written in a context, so totals are usually reported with no value for that dimension.
Do EBA and EIOPA filings use segment or scenario?
The EIOPA and Bank of England public samples Line4 counted put all dimension values in xbrli:scenario and none in xbrli:segment. Always follow the taxonomy, since each all arc fixes the element.
How do dimensions relate to the data point model and table linkbase?
In a data point model each cell of a template is a concept plus dimension values. The Table Linkbase, a Recommendation with errata to 17 December 2024, lays those cells out as the rows, columns and sheets firms see.
How are dimensions used outside Europe?
US listed companies use them in every Inline XBRL filing to the SEC, which accepts the 2026 and 2025 US GAAP taxonomies. One example from the SEC’s own guide: facts that apply only to eliminations between subsidiaries carry the ConsolidationItemsAxis with ConsolidationEliminationsMember. At the SEC the members go in the context’s segment, as scenario is not allowed, while EBA and EIOPA filings put them in scenario. Line4 validates both conventions, so a regulator can take dimensional data from firms whatever taxonomy they report with.
- XBRL International: Dimensions 1.0 (Recommendation 18 September 2006, errata to 25 January 2012)
- XBRL International: Table Linkbase 1.0
- EIOPA: Supervisory reporting, DPM and XBRL
- EBA: Reporting framework 4.2
- SEC: 2026 XBRL taxonomies update (March 2026)
- SEC: EDGAR XBRL Guide (November 2026), section 6.2.6
