Docs.rs

Name, type and overload resolution, subquery binding, and the bound logical plan.

Schemas, tables, views, constraints, dependency tracking, dictionaries and symbol tables.

The scalar, aggregate and window function library.

Fused kernels, Cranelift lowering and the compiled code cache.

MVCC versions, the write-ahead log and checkpointing.

The CSV reader and writer, including dialect sniffing and type inference.

Apache Iceberg table reads, v2 and v3.

Delta Lake table reads, through the kernel protocol.

Blitz networking

Reading and writing DuckDB storage format, at every version we claim to support.

The Parquet reader and writer, with page skipping and dictionary passthrough.

Blocks, row groups, column chunks, statistics and the buffer manager.

The generated cross product of operator, physical form and type, with runtime SIMD dispatch.

The JSON reader and the backing representation for the JSON functions.

The typed SSA expression IR that all four execution tiers share.