An MVCC-like columnar table on S3 with constant-time deletes

Parquet is excellent for analytical workloads. Columnar layout, aggressive compression, predicate pushdown, but deletes require rewriting entire files. Systems like Apache Iceberg and Delta Lake solve this by adding metadata layers that track delete files separately from data files. But what if, for fun, we built something (arguably) simpler? S3 now has conditional writes (If-Match, If-None-Match) that enable atomic operations without external coordination. Let’s explore how we might build a columnar table format on S3 that gets most of Parquet’s benefits while supporting constant-time deletes.