Bottom-up evaluation of DataHiLog

Many queries which may be useful in applications such as “data mining” and knowledge acquisition in databases cannot be expressed in traditional query languages such as SQL or Datalog because they lack the ability to query information regarding the schema of the database. We study the language DataHiLog introduced by Ross as a restriction of the language HiLog developed by Chen, Kifer and Warren. DataHiLog is still able to mix schema-level and instance-level queries, while avoiding the problem of infinite answers which can arise in HiLog. We show how the standard techniques for Datalog can be extended to provide a bottom-up evaluation mechanism for DataHiLog, and consider whether Datalog optimization techniques can also be extended to DataHiLog