The information retrieval R-tree (IR2- tree) which is the state of the art for answering the nearest neighbor queries defined explains an alternative solution based on the inverted index. the IR2-tree combines the Rtree with signature files. Next, we will review what is a signature file before explaining the details of IR2-trees. Our discussion assumes the knowledge of R-trees and the best-first algorithm for NN search, both of which are well-known techniques in spatial databases. Signature file in general refers to a hashing-based framework, whose instantiation is known as superimposed coding (SC), which is shown to be more effective than other instantiations Our treatment of nearest neighbor search falls in the general topic of spatial keyword search, which has also given rise to several alternative problems. A complete survey of all those problems goes beyond the scope of this paper. Below we mention several representatives, but interested readers can refer to for a nice survey. A form of keyword-based nearest neighbor queries that is similar to our formulation, but differs in how objects’ texts play a role in determining the query result. Specifically, aiming at an IR flavor, the approach of computes the relevance between the documents of an object p and a query q. This relevance score is then integrated with the Euclidean distance between p and q to calculate an overall similarity of p to q. The few objects with the highest similarity are returned. In this way, an object may still be in the query result, even though its document does not contain all the query keywords. In our method, same as , object texts are utilized in evaluating a boolean predicate, i.e., if any query keyword is missing in an object’s document, it must not be returned. Neither approach subsumes the other, and both make sense in different applications.