GIN Index with Array
Introduction GIN GIN (Generalized Inverted Index) also uses a B-tree data structure, so it shares similar characteristics with B-Tree. However, unlike B-tree, which stores each value along with a Tid to a single row in the main table, GIN stores the value along with information to access multiple rows sharing that same value (referred to as a Posting List). Therefore, the difference in usage is that B-tree is used to search for a row containing a specific value, while GIN is used to search for small values nested inside a row. Posting List You can understand it as a list containing IDs, which are primary keys corresponding to each row in the main table. If the table does not have an ID, it will use a physical identifier called ctid (Column Tuple Identifier) created by Postgres, with a structure consisting of (Page number, row index within the Page). Based on this ctid, Postgres can access any row in the main table. However, as you know, data in the heap is arranged arbitrarily and when...