Java JPA 和 Java JDBC 对同一个 MySQL 查询返回不同的结果

Question

注意：我知道这里有很多代码，但您可能只需要查看前 2 个代码块。

我有一个查询，根据该查询的结果，我构建了一个 ArrayList 的 Station 对象 ArrayList<Station>。

我想使用 JPA，因为从长远来看它更灵活，所以我正在调整我的代码以在 JPA 上运行。

所以我给一个方法提供了一个 Station 对象列表，如果该站满足特定的计算标准，这个方法将做一些计算并做station.setPingPong();，与问题无关。

问题是 JPA 的列表源自 3871 PingPongs，而 JDBC 的列表源自 3888 PingPongs，正确的是 3888。为什么会这样？

这是我使用数据库中的数据构建对象的代码：

for (int i = 0; i < uniqueStations.size(); i++) 
{

ArrayList<Station> stationDataList = new ArrayList<>();
ArrayList<Station> auxStationDataList = new ArrayList<>();
ArrayList<Station> pingPongProccessedData = new ArrayList<>();
ArrayList<Station> pingPongProccessedDataJPA = new ArrayList<>();

// Query using JDBC
String queryToApi = "SELECT * "
                    + "FROM " + readTable + " "
                    + "WHERE calling_station_id = '" + uniqueStations.get(i) + "' "
                    + "order by timestamp-acct_session_time, timestamp;";

// Query using JPA                    
String query = "SELECT timestamp-acct_session_time, timestamp, called_station_id "
                + "FROM " + readTable + " "
                + "WHERE calling_station_id = '" + uniqueStations.get(i) + "' "
                + "order by timestamp-acct_session_time, timestamp;";

// Using JDBC, here Station objects are created inside getStationData method, inserted into the list and returned.
stationDataList = Database.getInstance().getStationData(api.getConnectionSourceServer(), query, false);

// Using JPA, JavaLogs is my Class Entity from the database (it represents my table and was generated by NetBeans)
List<JavaLogs> javaLogsList =  api.executeSelectQuery(queryToApi, JavaLogs.class, "TesteV2PU");
int repetitionSTA = 0;
// Now I'm creating the Station objects                    
for(JavaLogs javaLogs : javaLogsList)
{
    int id = javaLogs.getIdacesso();      
    long timestamp = javaLogs.getTimestamp();
    long accountSessionTime = javaLogs.getAcctSessionTime();
    int startTime = (int) (timestamp - accountSessionTime);
    int endTime = javaLogs.getTimestamp();
    String accessPoint = javaLogs.getCalledStationId();
    int ttPrevious = 0;
    int ttNext = 0;
    int pPong = 0;
    Station station = new Station(id, startTime, endTime, accessPoint, stationAnonymized++, repetitionSTA++, ttPrevious, ttNext, pPong);

    auxStationDataList.add(station);
}

// The problem comes here, pingPongProccessedData does have the correct computation of Station objects
// But pingPongProccessedDataJPA doesn't have the correct computation of Station objects
// In other words, pingPongProccessedDataJPA is WRONG and pingPongProccessedData is correct.

pingPongProccessedData = detectPingPong(accessThreshold, transitionThreshold, stationDataList);
pingPongProccessedDataJPA = detectPingPong(accessThreshold, transitionThreshold, auxStationDataList);

pingPongStationList.addAll(pingPongProccessedData);
pingPongStationListJPA.addAll(pingPongProccessedDataJPA);

}

int totalPingPongsJDBC = 0;
int totalPingPongsJPA = 0;

for(Station station : pingPongStationList) 
{
    if(station.getPingPong() == 1)
    {
        totalPingPongsJDBC++;
    }
}

for(Station station : pingPongStationListJPA) 
{
    if(station.getPingPong() == 1)
    {
        totalPingPongsJPA++;
    }
}

// Here I get in the console: TOTAL PingPongs JPA: 3871 TOTAL PingPongs JDBC: 3888
System.out.println("TOTAL PingPongs JPA: " + totalPingPongsJPA + " TOTAL PingPongs JDBC: " + totalPingPongsJDBC);

这是来自我的Databse 对象的getStationData 方法：

public ArrayList<Station> getStationData(Connection con, String query, boolean isStationAnonymizeRequired) throws SQLException
{
    ArrayList<Station> stationList = new ArrayList<>();

    Statement s = con.createStatement();
    ResultSet rs = s.executeQuery(query);

    int repetitionSTA=1;

    while (rs.next()) 
    {
        /*
         * SQL Retorns:
         * StartTime - EndTime - AP - STA
         * 
         * Objecto Station:
         * ID - StartTime - EndTime - AP - STA - RepetitionSTA - TransitionTimePrevious - TransitionTimeNext - PingPong
         */

        if(isStationAnonymizeRequired == false)
        {
            Station station = new Station(id, rs.getInt(1), rs.getInt(2), rs.getString(3), stationMacAddress, repetitionSTA, 0, 0, 0);
            stationList.add(station);
            repetitionSTA++;
            id++;
        }
        else
        { //                            (id, startTime, endTime, ap, sta, repetitionSTA, ttprevious, ttnext, ppong)
            Station station = new Station(rs.getInt(1), rs.getInt(2), rs.getInt(3), rs.getString(4),
                    rs.getInt(5), repetitionSTA, rs.getInt(6), rs.getInt(7), rs.getInt(8));
            stationList.add(station);
            repetitionSTA++;
        }
    }
    rs.close();
    s.close();
    stationMacAddress++;

    return stationList;
}

我知道 ID 不同，构建方式不同，但这没有问题，因为 detectPingPong 方法中没有使用 ID。

以防万一，这是我的 detectPingPong 方法：

public ArrayList<Station> detectPingPong(int accessSessionTime, int transitionTime, ArrayList<Station> stationDataList)
{

    for(int i = 1; i<stationDataList.size()-1; i++) //i=1 to skip the first position and -1 to skip the last position
    {
        /*
         * StartTime - EndTime - AP - STA - TTprevious - TTnext - PPong - ID
         */

        /*
         * At current position values
         */
        int startTimeActual = stationDataList.get(i).getStartTime();
        int endTimeActual = stationDataList.get(i).getEndTime();
        String apActual = stationDataList.get(i).getAccessPoint();

        /*
         * get the values from the previous position
         * */ 
        int endTimePrevious = stationDataList.get(i-1).getEndTime();
        String apPrevious = stationDataList.get(i-1).getAccessPoint();

        /*
         * get the values from the next position
         * */
        int startTimeNext = stationDataList.get(i+1).getStartTime();
        String apNext = stationDataList.get(i+1).getAccessPoint();     


        //Calc transition times:
        int transitionTimePrevious = startTimeActual - endTimePrevious;
        int transitionTimeNext = startTimeNext - endTimeActual;

        stationDataList.get(i).setTransitionTimePrevious(transitionTimePrevious);
        stationDataList.get(i).setTransitionTimeNext(transitionTimeNext);


         /*
          * Testing the conditions to occur a pingpong
          * */
        if(transitionTimePrevious<0)
        {
            transitionTimePrevious=0;
            stationDataList.get(i).setTransitionTimePrevious(transitionTimePrevious);
        }


        if(transitionTimeNext<0)
        {
            transitionTimeNext=0;
            stationDataList.get(i).setTransitionTimeNext(transitionTimeNext);
        }


        /*
         * ***************************
         * *    TEST PING PONG       *
         * ***************************
         */


        if( endTimeActual - startTimeActual < accessSessionTime && 
                transitionTimePrevious < transitionTime &&
                transitionTimeNext < transitionTime &&
                apActual.equalsIgnoreCase(apPrevious) == false &&
                stationDataList.get(i).getRepetitionSTA() > 1 
           )
        {
            stationDataList.get(i).setPingPong();
        }

    }
    return stationDataList;
}

因此，由于两个列表的方法 (detectPingPong) 相同，并且我得到不同的结果，我假设错误在列表上，因此列表中的对象导致对象的构造。

Answer 1

您正在将结果映射到一个 bean 类 (javalogs)。可能有一些检索到的数据与JavaLogs 字段不匹配。那些检索到的行将被忽略。这就是为什么您会看到结果之间存在差异。