API Reference parquex v#0.3.0

Reads and writes Parquet files through reusable object stores.

A bounded columnar batch produced by a Parquex.Stream.

A validated Parquet dataset beneath one store prefix.

One successfully written dataset part.

A pull-based, prefix-pruned stream across dataset parts.

Summary of the parts written by one dataset writer.

A foreground, owner-bound writer for bounded time-partitioned Parquet parts.

A stable, contextual error returned by Parquex operations.

The ordered fields in a Parquet batch stream.

A named, typed, and nullable field in a Parquex.Schema.

A reusable local or S3-compatible object-storage namespace.

Metadata for one key in a reusable store namespace.

An owned, monitored staged object writer.

A single-pass, pull-based stream of bounded Parquex.Batch values.

Safe, bounded-cardinality telemetry helpers used by Parquex public operations.

The validated event-time partition contract for a dataset.

One canonical UTC partition and its half-open time interval.

An owned, incremental writer for one Parquet file.