Efficient low-cardinality ANDs in a search engine
问题 How do search engines such as Lucene, etc. perform AND queries where a term is common to many documents in the dataset? For example, in an inverted index of: term | document_id --------------------- program | 1, 2, 3, 5... python | 1, 4 code | 4 c++ | 4, 5 the term program is present in several documents meaning a query of program AND code would require performing an intersection upon a very large set of documents. Is there a way to perform AND queries without having to take the intersection