Module memory 

Constants

COMPRESSED_TO_DECODED_INFLATION 🔒

Compressed-to-decoded fallback used when the schema contains variable-width fields.

DATAFUSION_RUNTIME_FIXED_BYTES 🔒

Fixed task-context, operator, and allocator overhead observed for small streaming plans.

DATAFUSION_SORT_MERGE_RESERVATION_BYTES 🔒

DataFusion’s datafusion.execution.sort_spill_reservation_bytes default.

DATAFUSION_STREAMING_DECODED_WINDOW_BYTES 🔒

Upper bound for decoded Arrow data retained beyond the explicitly accounted record batches.

DATAFUSION_WRITER_WINDOW_BYTES 🔒

Output writer memory retained across input batches.

EQUALITY_DELETE_INFLATION 🔒

Compressed-size estimate for decoded equality-delete value buffers.

HASH_JOIN_ROW_OVERHEAD_BYTES 🔒

Per-row hash table and join bookkeeping floor for DataFusion’s retained HashJoinInput.

HEAP_FIXED_HEADROOM_BYTES 🔒

LARGE_SORT_THRESHOLD_BYTES 🔒

POSITION_DELETE_INFLATION 🔒

Compressed-size estimate for decoded position-delete value buffers.

POSITION_DELETE_KEY_OVERHEAD_BYTES 🔒

Per-row Arrow and hashing overhead for the two position-delete join keys.

SORT_TEMPORARY_HEADROOM_BYTES 🔒

STREAMING_PREFETCH_DECODE_OVERLAP_MIN_ACTIVE_FILES 🔒

At this scan concurrency, prefetched files and decoded buffers measurably overlap at peak.

Functions

complete_record_count 🔒

estimate_equality_delete_join_bytes 🔒

estimate_plan_memory

Estimates the peak heap bytes of one compaction plan for scheduler admission.

estimate_position_delete_join_overhead_bytes 🔒

estimate_prefetch_bytes 🔒

estimate_provider_prefetch 🔒

estimated_schema_row_width 🔒

hidden_row_width 🔒

position_delete_row_width 🔒

primitive_width 🔒

sum_file_sizes 🔒