如何使用 Spring Data Jpa 和 postgres 中的索引使我的本机查询更快

Question

I am trying to search a join table of millions of records using JPA native query and PostgreSQL but the result is too slow. I want to make it faster using index search on postgres or any better option that can be suggested. I looked up several articles online like this: https://hashrocket.com/blog/posts/exploring-postgres-gin-index

but am not quite sure how to implement it as am not getting the improved speed.

Here is what I had in my JPA repository:

 @Query(value = "SELECT DISTINCT C.APPROVED_NAME AS companyName, C.CITY AS city, C.STATE AS state, 
FROM TABLE_NAME_1 C LEFT JOIN TABLE_NAME_2 AS A ON A.COMPANY_FK = C.ID 
WHERE CONCAT(UPPER(A.FIRSTNAME), ' ', UPPER(A.SURNAME))
 LIKE UPPER(concat(:searchTerm,'%')) ", nativeQuery = true)
    List<table_1_Interface> findByName(@Param("searchTerm") String searchTerm);

Here is what I tried to do using index search (I am new to using index search):

 @Query(value = "CREATE INDEX company_search_idx ON TABLE_NAME_2(firstname, surname );
SELECT DISTINCT C.APPROVED_NAME AS companyName, C.CITY AS city, C.STATE AS state, 
FROM TABLE_NAME_1 C LEFT JOIN TABLE_NAME_2 AS A ON A.COMPANY_FK = C.ID 
WHERE CONCAT(UPPER(A.FIRSTNAME), ' ', UPPER(A.SURNAME))
 LIKE UPPER(concat(:searchTerm,'%')) ", nativeQuery = true)
    List<table_1_Interface> findByName(@Param("searchTerm") String searchTerm);

Please help me with your suggestions, ideas and solutions.

Answer 1

If you search on CONCAT(UPPER(A.FIRSTNAME), ' ', UPPER(A.SURNAME)) then that is what the index needs to be built on (but without the table aliases). But for pg_trgm it supports ILIKE directly, so you could just use that instead of manually fiddling with case.

Since you only add a % to the end of the search term, you could instead use a btree index with text_pattern_ops. But then you would need the UPPER with LIKE.