Spring-Data 预测。 正确的方法？

Question

In my attempt of using JPA with a legacy database, I've came to the conclusion I need to obtain a subset of my huge 50-60 column tables columns, since the queries generated by one single select sentence can be around 4 or 5 sheets of paper long with a endless number of joins.

I am wondering, how you experts, face this problem related mainly to performance and efficency, but also with the application's design and how you do structure your project, entities, pojos etc in a reasonable way.

• About the class containing the result:

  • Is it OK to use the full entity class to return 2 or 3 columns from a table?
  • Would you use a POJO in a DTO-fashioned way?
  • Or a shortened version of the entity referncing the same table?

• How would this interact with the dependent layers?

Answer 1

我会使用@Query 和自定义代码查询带有代表所需模型的 DTO，节省内存，也更喜欢 Spring-Data-JDBC 而不是 JPA

Answer 2

我会使用 Spring 的 JDBC 模板： https : //spring.io/guides/gs/relational-data-access/