Document query processing strategies: cost evaluation and heuristics

This paper describes query processing strategies used in a system where queries are specified on formatted data and text components of documents. The system provides different access methods: indexes for formatted data and signature files for the text. Four basic document queries, representative of a wide range of queries, are examined in detail. Strategies for query processing are presented together with an evaluation of the costs of the various strategies. Finally, general heuristics are derived.