the speed of querying directly on the client is very fast, but the speed in the code is much slower, and the amount of data is about 1 million

this is the query code:

    List<String> result = new ArrayList<>();
    String sql = "SELECT * FROM (SELECT KEYWORDS FROM KEYWORD_PERSONAL_HEAT WHERE USER_ID = ""+userId+"" AND KEYWORDS LIKE "%"+keyword+"%" GROUP BY KEYWORDS ORDER BY max(TIMES) DESC) WHERE ROWNUM <="+rowNum;
    List<Map> searchResult = JdbcUtils.findBySql(this.getEntityManager(), sql);
    for(Map map:searchResult){
        result.add((String) map.get("KEYWORDS"));
    }

this is the underlying query code:

public static List<Map> findBySql(EntityManager entityManager,String sql){
    List<Map> result=new ArrayList<Map>();
    Connection conn=null;
    PreparedStatement pstmt=null;
    ResultSet rs=null;
    try{
        conn=openConn(entityManager);
        pstmt=openPstmt(conn,sql);
        ResultSetMetaData rsmd=pstmt.getMetaData();
        rs=openRs(pstmt);
        while(rs.next()){
            Map data=new HashMap();
            for(int i=1;i<=rsmd.getColumnCount();iPP){
                String columnName=rsmd.getColumnName(i);
                data.put(columnName,rs.getObject(columnName));
            }
            result.add(data);
        }
    }catch (SQLException e){
        e.printStackTrace();
    }finally {
        closeConn(rs,pstmt,conn);
    }
    return result;
}