直接在客户端上查询速度很快，但是到了代码中速度慢了很多，数据量在100万左右

这是查询代码：

List 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 searchResult = JdbcUtils.findBySql(this.getEntityManager(), sql);

for(Map map:searchResult){

result.add((String) map.get("KEYWORDS"));

}

这是底层查询代码：

public static List findBySql(EntityManager entityManager,String sql){

List result=new ArrayList();

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();i++){

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;

}