Struct SchemaInferrer 

pub struct SchemaInferrer { /* private fields */ }

Automatic schema detector.

Samples column names and record values to produce a best-effort Schema. The inferred schema should be reviewed before use in production, call individual field overrides for any column the heuristics might misclassify.

Example

let schema = SchemaInferrer::new()
    .override_field("internal_code", FieldKind::Id)
    .override_field("notes",         FieldKind::FreeText)
    .infer(&records)
    .unwrap();

Implementations

impl SchemaInferrer

pub fn new() -> SchemaInferrer

Create a new inferrer loading heuristics from the embedded defaults (or from ZER_NAME_HEURISTICS / ZER_VALUE_PATTERNS env vars if set).

pub fn with_name_heuristics_file( self, path: impl AsRef<Path>, ) -> Result<SchemaInferrer, ZerError>

Override the name-based heuristics with rules loaded from a TOML file.

Returns Err if the file cannot be read or parsed.

pub fn with_value_patterns_file( self, path: impl AsRef<Path>, ) -> Result<SchemaInferrer, ZerError>

Override the value-pattern sampling with patterns loaded from a TOML file.

Returns Err if the file cannot be read, parsed, or contains an invalid regex.

pub fn override_field( self, name: impl Into<String>, kind: FieldKind, ) -> SchemaInferrer

Force a specific FieldKind for one field, bypassing inference.

This always takes precedence over both name-based and value-based heuristics.

pub fn infer(&self, records: &[Record]) -> Result<Schema, ZerError>

Infer a Schema from a sample of records.

50–100 non-null values per field is enough for reliable inference.

Returns Err(ZerError::EmptySchema) when records is empty (no field names can be discovered).

Trait Implementations

impl Default for SchemaInferrer

fn default() -> SchemaInferrer

Returns the “default value” for a type.