A method of automatically generating sql statements from human natural language for querying data

Abstract

The invention discloses a method for converting human natural language of query data into sql statement. The user enters a natural sentence, performs word segmentation on the sentence and removes stop words, and then performs automatic table name analysis. If the table name can be obtained, further processing is performed on the natural language after the word segmentation and removal of stop words and the obtained table name, otherwise Feedback prompt information after fuzzy matching; after obtaining the table name, judge whether multi-table association is based on metadata information, and if yes, perform query condition processing according to association rules; otherwise, directly perform single-table query condition processing; The language performs time analysis processing, and performs query field analysis, judgment condition analysis, cascade query analysis, and query aggregation analysis processing according to metadata information; finally, query sql is generated. The present invention does not require personnel to mark data, reduces the input of human resources, reduces the expenditure of labor costs, and solves the problem that the system cannot be reused caused by machine learning.

Description

technical field [0001] The invention relates to the field of natural language processing, in particular to a method for automatically generating SQL statements from human natural language for querying data. Background technique [0002] As the country's economy develops, more and more systems are put into use, generating a large amount of data. When non-professional users want to browse data, the traditional method usually presents the data in the form of charts in advance, and cannot query and browse data freely. The current semantic-to-sql method on the market usually collects a large amount of annotation information through machine learning, and generates SQL statements through long-term learning. This intermediate process requires personnel to label a large amount of data, and when new data or a new system is connected, it is usually marked again and a long period of machine learning is carried out. There are problems such as uncontrollable training time and unreusable ...

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to address the above problems and propose a method for automatically generating sql statements in human natural language for querying data. This method does not require personnel to label data and solves the problem of unreusable systems caused by machine learning.

Images