如何优化JPA查询

Question

I have a query which is taking a bit more time to fetch data from mysql db. There is only few records in database

@Query("    SELECT  COUNT(*)
    FROM  Message n
    WHERE  n.deleted = false
      AND  n.sent = true
      AND  n.senderEnvelope.user = ?1
      AND  n.replyToMessage IN (
        SELECT  DISTINCT m.replyToMessage
            FROM  Message m , Relationship r
            WHERE  m.deleted = false
              AND  m.sent = true
              AND  m.senderEnvelope.user = ?1
              AND  ?2 MEMBER OF r.categories 
                          )
       ")
Long countUniqueRepliesSentByUser(User currentUser ,
            RecommendationCategory recommendationCategory);

Is there is a way by which i can optimize this query or right it in another way. Please help. Thanks in advance

Answer 1

The biggest problem is that there is no relationship specified for these two tables:

FROM  Message m , Relationship r

That leads to a costly "cross join".

This seems like a syntax error; what was it supposed to mean??:

AND  ?2 MEMBER OF r.categories

After fixing those, then deal with IN . IN ( SELECT DISTINCT ... ) is poorly optimized. Furthermore the DISTINCT is, as some level, redundant. Consider changing it to

EXISTS ( SELECT 1 FROM ...
          AND n.replyToMessage = m.replyToMessage )

Please provide EXPLAIN SELECT ... for any query you are asking about.