Relational databases typically support vector search with functions such as dot product, cosine similarity, Euclidean distance and Manhattan distance. They use similarity scores to find K-Nearest ...