API reference

The public API is reached through two import surfaces: delta_engine.schema for declaring tables (no PySpark required) and delta_engine.databricks for building an engine and running syncs. Result and error types are re-exported from the top-level delta_engine package.

The full per-module reference is generated from the source tree; see delta_engine. The entry points:

You want to…

Start at

Declare tables, columns, keys

delta_engine.schema

Build an engine and sync

delta_engine.databricks

Inspect results and errors

delta_engine (top-level re-exports)

Notes on DeltaTable

scope (str, default "full")

Selects what the declaration manages. "full" manages the whole table. "metadata" restricts the sync to catalog metadata: comments, tags, and primary/foreign key constraints. "tags" restricts it to table and column tags. A restricted scope still declares the full table shape; aspects outside the scope are never changed, and any unmanaged drift causes validation to fail. Properties are the exception: a declaration that does not manage properties never compares them at all.

clustered_by (read-only accessor)

The tuple of liquid clustering key column names, in declaration order, reflecting the clustered_by constructor argument. A table-level list, the sibling of partitioned_by and mutually exclusive with it; at most four keys. Key order is not significant. See clustering.