Incomplete Answers for Queries over Semistructured Data.
Yaron Kanza, Werner Nutt, Yehoshua Sagiv · 1999
Semistructured data are inherently incomplete and their structure does not comply with a strict schema. A query language has to take this into account and allow answers to be incomplete. This paper presents a model of queries over semistructured data that allows for incomplete answers, i.e., answers in which not all the variables of the query are bound. We define a simple data model that is similar to OEM [PGMW95]. For this data model we present the concept of a query language based on constraints. The paper describes the different constraints that are used in the query and the different phases needed for evaluating them. We propose the concept of maximal answers that are obtained by matching a maximal portion of the database graph with the query constraints. Maximalmatchings and constraints that can be evaluated over partial data give us the ability to retrieve partial answers when complete answers do not exist. 1 Introduction Semistructured data lack a homogeneous structure and are ...