All posts published on this site.
Everything published here, newest first.
DataFusion evaluates a physical expression tree by walking it once per RecordBatch, dispatching on a Rust enum at each node, and handing whole arrays to Arrow kernels — never …
DataFusion validates the plan after every optimizer rule, using a two-level invariant system where the cheap structural checks always run and the expensive semantic checks are …
DataFusion’s hash join builds one hash table per partition: a hashbrown table maps a folded 64-bit hash to a chain of row indices stored in a flat u32/u64 array, and the payload …
DataFusion splits group-key interning into a hash table of (hash, group id) plus one column builder per key column, so multi-column GROUP BY never materializes a row per group.
ClickHouse charges every allocation to a chain of memory trackers — global, user, query, thread.