Module index 

Per-document structural index for O(1) field access.

When a query accesses multiple fields from the same document (e.g., GROUP BY + aggregate + filter), building a FieldIndex once and reusing it for all field lookups avoids repeated O(N) key scanning.

Uses a flat array with linear search for small docs (≤ 16 fields) to avoid HashMap allocation overhead. Falls back to HashMap for large docs.

Structs

FieldIndex

Pre-computed field offset table for a single MessagePack document.