Expand description
Data structures for Kermit’s relational algebra engine.
Provides two trie-based relation implementations:
TreeTrie: A pointer-based trie where each node owns its children. Simple and cache-friendly for small relations.ColumnTrie: A column-oriented (flattened) trie that stores each level in paralleldata/intervalarrays. More compact for large relations.
Both implement the Relation and
TrieIterable traits, making them
interchangeable in join algorithms.
Structs§
- Column
Trie - A column-oriented trie that stores a relation as parallel arrays per level.
- Relation
Header - Metadata for a relation: its name, attribute names, and arity.
- Tree
Trie - A pointer-based trie that stores a relation as a tree of
TrieNodes.
Enums§
- Index
Structure - The available trie-based index structures for storing relations.
- Relation
Error - Error type for relation file operations (CSV and Parquet).
Traits§
- Heap
Size - Trait for calculating heap-allocated memory usage.
- Projectable
- A relation that can produce a new relation containing only the specified columns.
- Relation
- The
Relationtrait defines a relational data structure that can store and retrieve tuples ofusizekeys, and participate in join operations. - Relation
File Ext - Extension trait for
Relationto add file reading capabilities.