정보검색을 위한 TWO LEVEL MAPPED RANDOM FILE

File structures and the factors that affect the performances of file organizations are investigated. In general, the list structured file structures accommodate the multiple and generic keyed records. The shortcomings of the list structured file structures are the retrieval time and storage efficiencies when the amount of data becomes extremely large. A two level mapped random file, named PIS, is developed and tested its performance using the developed experimental information retrieval system, ADAM. In this system, a set of records can be retrieved with the anthor key(s) and/or the subject key(s). The experiments reveal that the performance of the PIS is superior to the other file structures.

특수 형태의 2-level mapped random file(PIS라 불리움)을 제시하며 이 개념을 CDC CYBER73 컴퓨터에 구체적으로 구현시켜서 같은 컴퓨터 내에서의 inverted list file 구조로 사용가능한 유일한 file로서의 SCOPE Indexed Sequential(SIS) file과 성능과 비용을 분석하였다.