How to do group by substring in Eclipselink?

问题

I'm trying to group the store by store country and see how many stores there are in each country. The store country happens to be the first 2 characters of the store's code. That means if the store code is "US000010", the country is "US"

My entity object:

...
public class Store {
    @column(name = "code")
    private String code;
    ...    
}

My JPQL is:

  SELECT substring(s.code, 0, 2),
         count(s) 
    FROM Store s 
GROUP BY substring(s.code, 0, 2)

This keeps throwing me ORA-00979: not a GROUP BY expression, which I suspect is due to the JPQL above that is converted to

SELECT SUBSTR(CODE, ?, ?), COUNT(CODE) FROM Store GROUP BY SUBSTR(CODE, ?, ?)

Excluding other alternatives like creating a view or using native query, is there any way I perform a group by substring in Eclipselink using JPQL?

回答1:

I recently had a similar problem grouping by a TO_CHAR function. The solution I found is ExclipseLink specific, and it was just to set a hint on the query: query.setHint(QueryHints.BIND_PARAMETERS, HintValues.FALSE);.

Put this code before executing the query and it will not use parameters, but replace the values directly on the SQL string sent to the JDBC driver

来源：https://stackoverflow.com/questions/34433648/how-to-do-group-by-substring-in-eclipselink