Expand description
§faucet-sink-delta
Apache Delta Lake sink for the faucet-stream
ecosystem. Appends JSON records to a Delta table on the local filesystem or
cloud object storage (S3 / Azure / GCS) via the Rust
deltalake crate — the idiomatic,
high-throughput way to land data for Databricks (and Spark, Trino, DuckDB,
Microsoft Fabric) at the open table-format level.
- Lazily creates the table from the inferred schema on first write
(
create_if_not_missing), honouringpartition_by. - Appends one atomic Delta commit per
flush. - Append-only in v1 —
Sink::supported_write_modesreturns[Append](MERGE/upsert is a version-gated follow-up). - No datafusion dependency: append uses delta-rs’s low-level
RecordBatchWriter.
Cloud backends are opt-in cargo features: s3, azure, gcs.
Structs§
- Delta
Connection - Location + credentials for a Delta table. Flattened (
#[serde(flatten)]) into both the source and sink config so the same three keys appear at the connector-config top level. - Delta
Sink - A sink that appends JSON records to an Apache Delta Lake table.
- Delta
Sink Config - Configuration for the Delta Lake sink connector (append-only in v1).
Enums§
- Delta
Credentials - Object-store credentials for reading/writing a Delta table.
Constants§
- DEFAULT_
SAMPLE_ SIZE - Default number of records sampled to infer the table schema on first write.
- DESCRIPTOR
- The connector’s static descriptor (RFC 0013): its capabilities, answered from config alone.