Automatic statement conversion method and device

Abstract

An automatic statement conversion method divides an original SQL statement into statement sequences, and comprises: a preprocessing step, i.e., sequentially analyzing and obtaining semantic values ofan INSERT statement and a UPDATE statement, and respectively storing analysis results in a preset storage structure; obtaining an operation target, a condition and assignment information of a simple UPDATE statement, and filling a merged storage position of the INSERT statement with the same operation target as the UPDATE statement with the condition and assignment information of the UPDATE statement; a conversion step, i.e., sequentially converting and outputting the statement sequences subjected to the preprocessing step; directly outputting the CREATE statement and combining the INSERT statement with the null content in the storage position; updating and outputting the INSERT statement in the merged storage position; and outputting the complex UPDATE statement subjected to conversion. The invention further discloses an automatic statement conversion device. According to the method and the device, the UPDATE statement is automatically converted into the INSERT statement.

Description

technical field [0001] The invention relates to the software industry, and more specifically, relates to a sentence conversion technology. Background technique [0002] With the continuous expansion of network scale and rapid data growth, traditional data processing methods based on relational databases have become more and more inadequate. The most economical and smooth way to transition to big data technology is to directly reuse a series of SQL statements written in the original data processing process, which can ensure that the original business logic remains unchanged, but only changes the underlying execution engine. [0003] The biggest problem in the migration process is that the current big data processing-oriented databases or computing frameworks (such as Greenplum and Spark) are all analysis-oriented, and the support for the update operation in the original processing logic is very weak or not supported. It is necessary to provide an automatic conversion tool t...

AI Technical Summary

Problems solved by technology

[0008] For the above scheme a), since the current big data processing-oriented databases or computing frameworks (such as GreenPlum and Spark) are all analysis-oriented, the support for the update operation in the original processing logic is very weak or does not support, if the existing If there is an update statement in the SQL sequence, the efficiency will be very low or it will not be supported directly

[0009] The above scheme b) is a manual transplantation scheme, which has a large workload of transplantation (one SQL has hundreds of fields), and it is easy to cause logic errors. The workload of data verification and comparison between the old and new systems is very huge

Images