X2S: Translating XPath into Efficient SQL Queries

Translating XPath into SQL queries is necessary if XML documents are stored in relational databases. Although most of the current methods can translate XPath into equivalent SQL queries, they seldom focus on the efficiency of the translated SQL queries. This paper proposes a method called X2S to handle the problem. X2S is based on the extended relational encoding, which can deal with the operators in XPath efficiently. X2S also exploits DTD to rewrite the XPath logically before the SQL translation. The goal of the rewriting is to eliminate the wildcards of the XPath, to remove the ancestor/descendant relationship and to shorten the length of the XPath. An estimation cost model is used to choose one with the lowest cost from the candidate SQL statements. The experimental results demonstrate that the SQL queries generated by X2S outperform those from other methods significantly.