Index store management method based on zipper information

Abstract

The invention discloses an index store management method based on zipper information. A physical file organization of index comprises a B tree index file, a zipper file and a data file; the index store management method based on zipper information comprises the following steps: writing data to the data file according to key value pairs; constructing a zipper information block according to the written data information; and judging whether inserted data key value pairs exist on the B tree, and accordingly updating a B tree file and the zipper file. The zipper file is added to the physical file design, in logic process of index management, data operation is conducted based on zipper information items, so that support for effective inserting, updating and inquiring of the data is provided, and data process performance is improved.

Description

technical field [0001] The invention relates to the fields of database and information retrieval, in particular to an index storage management method based on zipper information. Background technique [0002] The emergence of the Web has greatly promoted the research on database systems. At the same time, various technologies in the field of general information retrieval have also been reborn in Web information retrieval. The research on information retrieval, database systems and the Web itself is being integrated in various ways. How to quickly find information that meets user needs from a large amount of stored data has become a new hotspot in the field of database and information retrieval. [0003] In database systems, indexes are usually used to assist in fast query and improve the speed of database table data access. Generally speaking, the basic unit of database storage is a page, and the disk space allocated by the physical file for storing information can be logic...

AI Technical Summary

Problems solved by technology

[0011] However, the current index storage scheme also has some defects. The main reason is that when updating the index, the key corresponding to the inserted data may already exist in the B-tree. If you directly update the physical data file, you need to find the corresponding key first. The position of the data block in the file, and then append and write data inside the file

However, the current file writing operation only supports appending at the end of the file. Therefore, this operation needs to rearrange the existing data pages in the file and write the newly added data to the appropriate position, resulting in more file fragments and reducing data loss. Update and query performance

Excessive file fragments will cause the system to search back and forth when reading files, causing system performance to degrade, and in severe cases, it may also lead to loss of stored files

Images