Hi, I’ve got an application where I’d like to search a collection of objects based on various properties, some text and others non-text (bools, enums, ints, etc). I’ve used full text search before, following the PG docs to set up a index on a ts_vector. And of course I’ve used normal indexes before for accelerating basic queries that use non-text columns. Any recommendations on combining the two? For example, if I search for objects where the description text matches [some ts_query] and the color = red, I can imagine putting the color property into the text index somehow - maybe with tokens like: ’color_red’, ‘color_blue’, but for something like an integer … ? Rob
