This document discusses spatial approximate string search in large spatial databases. It proposes the MHR-tree and RSASSOL methods. The MHR-tree embeds min-wise signatures into an R-tree to enable approximate string search in Euclidean space. RSASSOL combines q-gram inverted lists and reference node pruning for exact string search on road networks. Experiments on real datasets with millions of points and hundreds of thousands of nodes show the efficiency and effectiveness of the proposed approaches over baseline methods.