Preference SQL{1}

by Irving A

Customer’s wishes and preferences are very important when shopping online. These preferences are becoming extremely important in search engines and SQL preference design. Customer tastes are classified into two categories: knock out criteria that must be fulfilled versus soft criteria that should be fulfilled as closely as possible. During physical shopping, a customer expects a sales person to assist in finding the most suitable item. The same should occur during online shopping; user preferences during e-shopping sessions. When users search for an item and the results return unmatched, oftentimes customers leave the page frustrated and unhappy. This unpleasant system behavior has encouraged more efficient customer preference system design. Most systems are overloaded with lots of mostly irrelevant information. Although many approaches have been used to resolve the empty results problem, such as parametric search, techniques of query relaxation or case-based reasoning have all failed at solving the issue.


Preference SQL language design is an extension of the standard SQL design supporting a bundle of built-in base preferences types and further operators to build more complex preference types. Search engines involving only standard SQL suffer from understanding the notion of preferences and are incapable of directly supporting soft constraints. The model for Preference SQL design has two faces: choices in an “exact world” and choices in the “real world.” In the first, all user wishes are satisfied completely or not at all. In technical terms, the items are either exact matches or not. In the second face the choices are guided by personal preferences. This system is geared towards match-making, which finds the best possible match between one’s wishes and the reality, or will lead to the notion of soft constraints.

Successful personalized search engines should have hard constraints in an exact match world and preference-driven choices. Preference SQL language aims at becoming a suitable model for preference-driven search results. This system expands online sales and speeds up the process at finding the perfect products that best match our tastes. At times these preferences are stored based on our “favorites” or based on products we purchase frequently. Many online stores already use this type of personalized search language. When results return as empty, customers will stop searching online for items and return to in-store physical shopping. After all, a salesperson executes the job of a preference SQL search engine in the Internet world.


KieBling, W. (2002). AMC DIGITAL LIBRARY. Retrieved November 2009, from CSPU Pomona: