executeWritePipeline
The main entry point for persisting mutation data through the write pipeline.
Load Strategies
How full_load, full_load_append, and append strategies control data storage and table management.
Partitioning
How data is partitioned into Hive-style paths within the all.parquet/ directory.