The database-driven CRN has emerged as a promising solution for the spectrum scarcity issue. However, it also raises severe privacy concerns. Although there are some existing works on this topic, they are far from practical due to their restrict on particular database structures or extremely heavy computation and communication overhead. To address these issues, in this paper we develop two practical privacy-preserving spectrum query schemes. A basic scheme is first proposed based on private equality test (PET) technique. In order to enhance its efficiency, an advanced scheme is further devised by integrating the Locality Sensitive Hash (LSH). More importantly, both of them are applicable to multi-SP scenarios, which have never been discussed previously.