Abstract: Old spatial queries, like range search, nearest neighbour retrieval, involved only conditions on objects’ geometric properties. gradually many new applications call for novel forms of queries which aim to find objects which satisfy both a spatial predicate, as well as a predicate on their associated texts. consider an example: instead of considering all the hotels, a nearest neighbour query would just ask for the hotel that lies closest among the ones whose menu cards contain “rolls, burger, caramel custard” all at the same moment. at this time the best possible solution to such type of queries is based on the ir²-tree, which, as shown here, has some deficiencies that seriously would impact its efficiency. inspired by this, we are developing a new method called the spatial inverted index which extends the conventional inverted index to cope with multidimensional data, and produces algorithms which will answer nearest neighbor queries with keywords in real time.

Keywords: ir2 tree, spatial inverted index , r-tree.