Define Your Data

Data ModelSQLSchema

AetheriusDB is a strict, typed relational engine. Every table has a declared schema, every column has one precise type, and the standard ANSI SQL you already know works against it — plus a handful of native extensions for time-travel, storage residency, and workload-aware optimization.

This section covers how you define and evolve your data: declaring tables and types, the SQL surface, versioning schema like code, introspecting the catalog, declaring your workload’s shape, and deploying your own compute logic.

Why it matters

Strict typing and a queryable catalog aren’t red tape — they’re what let the engine run your data fast and safely. A column has one precise type, so there’s no per-row guessing at read time; the catalog is just tables, so your schema is introspectable and versionable with the same SQL you already write.

• Correct by construction — declared types are enforced at write time, so bad data is rejected at the door instead of corrupting a query months later.
• Evolve your schema like code — branch it, migrate it, and roll it back the way you branch application code.
• Standard SQL, no new dialect — the ANSI surface you know works as-is; the native extensions (time-travel, residency) are additive.

How fast it is

• Instant schema changes. ADD COLUMN … DEFAULT is a metadata write, not a table rewrite — it returns immediately no matter how large the table.
• Millisecond rollback. Reverting a migration repoints the catalog; there’s no dump-and-reload.

Pages in this section

• Relational Data Model — strictly-typed tables, measure and dimension kinds, and why strict typing pays off.
• ANSI SQL Compliance — the standard SQL surface plus native DDL extensions like AS OF TIMESTAMP.
• Schema-as-Code & Migrations — instant DDL, Git-style branching, and millisecond rollback.
• Catalog & Metadata — the catalog as ordinary queryable tables, with full schema history.
• Cognitive Schema — declare your workload’s shape so the engine prepares for it ahead of time.
• Functions & Procedures — deploy compiled WebAssembly compute cells instead of stored-procedure source.
• Hardware-Aware Semantic Types — a Beta family of specialised column types for trends, tags, geo, and IDs.

Each feature page carries an availability badge (Stable / Opt-in / Beta / Roadmap). The authoritative flag list lives in the configuration reference.

SQL Grammar & Types →

The exact type system, CTEs, window functions, and NULL semantics.

Quickstart →

SQL, a vector query, and a WASM cell end-to-end.

Configuration & Availability →

Which features are on by default vs. opt-in.