Module materializer 

rbt::materializer: multi-format writers including filesystem Iceberg-style tables.

Streaming materialize (stream) is the default engine path: pull a DataFusion batch stream, write batch-by-batch, drop each batch, atomic-publish the file.

Re-exports

pub use iceberg_catalog::verify_iceberg_catalog_table;

pub use iceberg_catalog::write_iceberg_catalog_batches;

pub use iceberg_catalog::write_iceberg_catalog_stream;

pub use iceberg_catalog::IcebergCatalogOptions;

pub use iceberg_catalog::IcebergCatalogWriteStats;

pub use stream::atomic_publish;

pub use stream::load_parquet_batches;

pub use stream::materialize_stream;

pub use stream::partial_path_for;

pub use stream::write_empty_parquet;

pub use stream::write_parquet_batches_atomic;

pub use stream::write_parquet_stream;

pub use stream::MaterializeWriteOptions;

pub use stream::StreamWriteStats;

Modules

iceberg_catalog

Official Apache Iceberg catalog materialize (P2 proof gate).

stream

Streaming materialize: pull DF RecordBatch streams batch-by-batch, write, drop.

Structs

MultiFormatWriter

Multi-format stream writer (Parquet, JSONL, CSV, filesystem Iceberg layout).

WapMaterializer

Write-Audit-Publish (WAP) Materializer — audit helpers (catalog branch publish later).

Enums

WapStatus

Status of a Write-Audit-Publish materialization run.

Functions

sibling_iceberg_dir

Sibling Iceberg table directory for dual-write: foo.parquet → foo.iceberg/.

write_iceberg_fs_table

Write an Iceberg-style filesystem table (data files + metadata snapshot).