Date Added: Oct 2010
The authors present a structured approach to SPARQL to SQL translation using AQL - a purpose-built intermediate query language. The approach produces a single SQL query for a single SPARQL query. Using AQL, they revisit the semantic mismatch between SPARQL and SQL and present query transformations on AQL presentation which enable the correct translation of some difficult corner cases. By using explicit expression type features in AQL, they also present type inference for expressions. They demonstrate the benefit of type inference as a basis for semantically correct optimizations in translation.