Hey everyone,
I’d like to share the design doc for IcebergIO’s new CDC (merge-on-read) sink, which applies a stream of inserts, updates, and deletes to an Iceberg table: https://s.apache.org/beam-iceberg-cdc-sink It covers the high-level design of the sink and compares it with the Flink equivalent. Together with the Iceberg CDC source <https://s.apache.org/beam-iceberg-incremental-cdc-source> (released in 2.76), a Beam pipeline can now keep one Iceberg table in sync with another. This is available in both batch and streaming. Questions and feedback are welcome! Thanks, Ahmed
