Module schema 

Schema registry with schema-on-write and additive evolution.

The schema registry tracks the column definitions for each measurement. Schemas are created automatically on first write and evolve additively: new fields/tags are added automatically, but type changes are rejected.

Schema Evolution Strategy

Chronix uses a schema-on-write approach to schema evolution:

• Automatic creation: The first write to a measurement creates the schema with the timestamp column plus all tags and fields from that point.
• Additive evolution: Subsequent writes may introduce new tag or field columns. These are appended to the schema automatically — no DDL statement is required.
• Type safety: If a field name already exists with a different type, the write is rejected with SchemaError::TypeConflict.
• Durability: the catalog manifest is the only durable record of a schema; the database persists every action register_batch returns before it writes the data that needs it
• Query-time safety: Queries that reference a column not present in older segments return zero rows for those segments (no panic).

Current Limitations

• No column removal: Once a tag or field is added, it cannot be dropped. Obsolete columns simply receive null values in new writes.
• No column rename: Renaming requires writing a migration that copies data from the old column to a new one and stops writing the old name.
• No type change: A field’s type is immutable once established. To change a field from i64 to f64, create a new field name.
• No column reorder: Columns are appended in arrival order.

Structs

ColumnDef

A column definition within a measurement schema.

MeasurementSchema

Schema for a single measurement (table).

SchemaRegistry

Thread-safe, in-memory schema registry.

Enums

ColumnRole

The role of a column within a measurement.

ColumnType

The data type of a column.

SchemaAction

A schema action that can be applied to the registry.