一种对外存快速搜索的文件结构

A Fast Retrieval File Organization

对于搜索操作使用频繁的文件系统(如目录索引文件),本文提出一种新的文件结构,能够保证只需一次盘操作就能搜索到所需查询的记录。它是利用在核心空间开辟一数组——分离器数组来进行搜索的。这个数组所占内存一般每块只需几位(UNIX 操作系统中每块为512字节)。用此方法改进后的 UNIX 文件系统在搜索某一文件路径名分量时能直接确定在哪一盘块上。

Retrieval operation is often used in some file system (eg.direct file).This paper introduces a new file organization method which guarantees retrieval of any record in one access.This method is based on hashing and makes use of an array in core, called separator,to direct the search.The array is very small,a few bits for each block (512 bits/block in UNIX).The improved UNIX file system can determine which block the message is stored immediatelly according to path-name.