A Method for Realizing Final Consistency Full-Text Retrieval in High Performance Database

Abstract

The invention discloses a method for implementing ultimate consistency full-text search of a high-performance database. A search engine is connected with the database through a connector; by establishment of a temporary circular log file of the database, under the assistance of a tool in the database, data contents in the search engine are always kept consistent with contents of fields needing to undergo full-text search in the database, accuracy in the search process is ensured, retrieval efficiency is improved, and a problem of high development and maintenance difficulty caused by direct use of an inverted index mode in the database and a shortcoming of poor stability caused by a conventional double-write mode are avoided. The method has the advantages that the data contents are complete, the establishing process is simple and the method is convenient to use; and the method is particularly suitable for data retrieval at a cloud.

Description

technical field [0001] The invention relates to a database retrieval method, in particular to a full-text retrieval method capable of realizing the final consistency of a high-performance database. Background technique [0002] In the world, there are many different technical products for new distributed databases, such as MongoDB, Cassandra, HBase, etc. In China, it includes SequoiaDB giant sequoia database, Alibaba's OceanBase, etc. The core functions of these databases are the storage and retrieval of records. [0003] Generally speaking, the precise search mechanism of the database includes operations such as equal to, greater than less than, greater than or equal to, less than or equal to, not equal to, etc., while fuzzy queries generally support the LIKE mechanism of SQL or regular expressions. In terms of execution efficiency, the efficiency of fuzzy query is far lower than that of precise retrieval with well-established indexes. This is because most databases use t...

AI Technical Summary

Problems solved by technology

Although this approach is simple and convenient, it may cause inconsistency in stored data, especially in concurrent scenarios, it is easy for multiple tasks to change the same record, resulting in inconsistencies between search engines and databases

In the worst case, if the application server crashes after writing to the database and does not have time to write to the search engine, the search engine will not be able to retrieve this record, resulting in retrieval failure

Another form is to implement the inverted index directly inside the database, but the difficulty of development and maintenance is greatly increased, and the cost and complexity of later use are high, and the popularity is poor