Module schema 

Schema inference engine for columnar JSON/NDJSON data.

Takes the output of ndjson::preprocess (columns separated by \x00, values within columns separated by \x01) and infers the type of each column by examining its values.

The inferred schema can be serialized into compact binary metadata for storage in .dcx transform metadata, and deserialized by the decoder.

Structs

ColumnSchema

Schema for a single column: type + null statistics.

InferredSchema

Inferred schema for an entire columnar dataset.

Enums

ColumnType

Inferred type of a single column.

TimestampFormat

Detected timestamp format.

Functions

deserialize_schema

Deserialize schema from transform metadata bytes.

infer_schema

Infer schema from columnar data (post ndjson::preprocess output).

serialize_schema

Serialize schema to compact binary bytes for transform metadata.