Question

我将使用JoinRowSet加入两个表（它们位于不同的DB服务器上）。我能够让它运作起来。但是，我发现了一个严重的问题。我的joines resultSet中有两行丢失。为了证明这一点，我加入了相同的查询，但仍然缺少两行（我认为我必须返回一个结果集，每个连接集的行数相同）。

这是我的代码：

@Stateless
public class TestEJB {

    public void getRektCrossJoin() throws Exception {
        Connection firstDB = null;
        Connection secondDB = null;

        try {
            Context ic = new InitialContext();
            try {
                firstDB = ((DataSource) ic.lookup("java:app/env/jdbc/firstDB")).getConnection();
            } catch (Exception e) {
                firstDB = ((DataSource) ic.lookup("java:comp/env/jdbc/firstDB")).getConnection();
            }

            try {
                secondDB = ((DataSource) ic.lookup("java:app/env/jdbc/secondDB")).getConnection();
            } catch (Exception e) {
                secondDB = ((DataSource) ic.lookup("java:comp/env/jdbc/secondDB")).getConnection();
            }

        } catch (Exception e) {
            System.err.println(e.getMessage());
        }

        String sqlAfa = "SELECT * FROM afa order by percent;";

        try (PreparedStatement stmtS = firstDB.prepareStatement(sqlAfa);PreparedStatement stmtJ = secondDB.prepareStatement(sqlAfa)) {

            RowSetFactory rowSetFactory = RowSetProvider.newFactory();
            try (ResultSet rsS = stmtS.executeQuery(); 
                    ResultSet rsJ = stmtJ.executeQuery(); 
                    JoinRowSet joinRowSet = rowSetFactory.createJoinRowSet()) {

                System.out.println("\n >>> First DB <<< ");
                int columnCountS = rsS.getMetaData().getColumnCount();
                for(int i = 1; i<= columnCountS; i++) {
                    System.out.print(rsS.getMetaData().getColumnName(i) + "\t");
                }
                System.out.println();

                while (rsS.next()) {
                    for(int i=1; i<= columnCountS; i++) {
                        System.out.print(rsS.getString(i) + "\t");
                    }
                    System.out.println();
                }


                System.out.println("\n >>> Second DB <<< ");
                int columnCountJ = rsJ.getMetaData().getColumnCount();
                for(int i = 1; i<= columnCountJ; i++) {
                    System.out.print(rsJ.getMetaData().getColumnName(i) + "\t");
                }
                System.out.println();
                while (rsJ.next()) {
                    for(int i=1; i<= columnCountJ; i++) {
                        System.out.print(rsJ.getString(i) + "\t");
                    }
                    System.out.println();
                }
                rsS.beforeFirst();
                rsJ.beforeFirst();

                CachedRowSet cachedRowSet = rowSetFactory.createCachedRowSet();
                CachedRowSet cachedRowSet2 = rowSetFactory.createCachedRowSet();

                cachedRowSet.populate(rsS);
                joinRowSet.addRowSet(cachedRowSet,2); //second field is unique

                cachedRowSet2.populate(rsJ);
                joinRowSet.addRowSet(cachedRowSet2,2); //second field is unique

                joinRowSet.first();
                if (!joinRowSet.next() ) {
                    System.err.println("No data");
                } else {
                    int columnCount = joinRowSet.getMetaData().getColumnCount();
                    for(int i = 1; i<= columnCount; i++) {
                        System.out.print(joinRowSet.getMetaData().getColumnName(i) + "\t");
                    }
                    System.out.println();
                    while(joinRowSet.next()) {

                        for(int i=1; i<= columnCount; i++) {
                            System.out.print(joinRowSet.getString(i) + "\t");
                        }
                        System.out.println();
                    }
                }

                cachedRowSet.close();
                cachedRowSet2.close();
            }
        }

        firstDB.close();
        secondDB.close();
    }

这是输出：

INFO [stdout] Southpark DB megvan
INFO [stdout] Jerry DB megvan
INFO [stdout] 
INFO [stdout]  >>> First DB <<< 
INFO [stdout] category  percent from_date   to_date 
INFO [stdout] 0 0   1900-01-01  2030-12-31  
INFO [stdout] 2 5   2018-01-01  2030-12-31  
INFO [stdout] 2 18  2017-01-01  2017-12-31  
INFO [stdout] 1 20  1900-01-01  2009-06-30  
INFO [stdout] 1 25  2009-07-01  2011-12-31  
INFO [stdout] 1 27  2012-01-01  2030-12-31  
INFO [stdout] 
INFO [stdout]  >>> Second DB <<< 
INFO [stdout] category  percent from_date   to_date 
INFO [stdout] 0 0   1900-01-01  2030-12-31  
INFO [stdout] 2 5   2018-01-01  2030-12-31  
INFO [stdout] 2 18  2017-01-01  2017-12-31  
INFO [stdout] 1 20  1900-01-01  2009-06-30  
INFO [stdout] 1 25  2009-07-01  2011-12-31  
INFO [stdout] 1 27  2012-01-01  2030-12-31  
INFO [stdout] 
INFO [stdout]  >>> JoinRowSet <<< 
INFO [stdout] category  percent from_date   to_date category    from_date   to_date 
INFO [stdout] 1 20.0    1900-01-01  2009-06-30  1   1900-01-01  2009-06-30  
INFO [stdout] 2 18.0    2017-01-01  2017-12-31  2   2017-01-01  2017-12-31  
INFO [stdout] 2 5.0 2018-01-01  2030-12-31  2   2018-01-01  2030-12-31  
INFO [stdout] 0 0.0 1900-01-01  2030-12-31  0   1900-01-01  2030-12-31

可能发生了什么？提前谢谢！

编辑：我复制了afa选择：

String sqlAfa = "SELECT * FROM afa order by percent DESC;";
String sqlAfa2 = "SELECT * FROM afa order by percent ASC;";

之后我修改了outter try-with-resources块：

PreparedStatement stmtJ = secondDB.prepareStatement(sqlAfa2)

这是输出：

INFO [stdout]  >>> First DB  <<< 
INFO [stdout] category  percent from_date   to_date 
INFO [stdout] 1 27  2012-01-01  2030-12-31  
INFO [stdout] 1 25  2009-07-01  2011-12-31  
INFO [stdout] 1 20  1900-01-01  2009-06-30  
INFO [stdout] 2 18  2017-01-01  2017-12-31  
INFO [stdout] 2 5   2018-01-01  2030-12-31  
INFO [stdout] 0 0   1900-01-01  2030-12-31  
INFO [stdout] 
INFO [stdout]  >>> Second DB <<< 
INFO [stdout] category  percent from_date   to_date 
INFO [stdout] 0 0   1900-01-01  2030-12-31  
INFO [stdout] 2 5   2018-01-01  2030-12-31  
INFO [stdout] 2 18  2017-01-01  2017-12-31  
INFO [stdout] 1 20  1900-01-01  2009-06-30  
INFO [stdout] 1 25  2009-07-01  2011-12-31  
INFO [stdout] 1 27  2012-01-01  2030-12-31  
INFO [stdout] 
INFO [stdout]  >>> JoinRowSet <<< 
INFO [stdout] category  percent from_date   to_date category    from_date   to_date 
INFO [stdout] 2 18.0    2017-01-01  2017-12-31  2   2017-01-01  2017-12-31  
INFO [stdout] 1 20.0    1900-01-01  2009-06-30  1   1900-01-01  2009-06-30  
INFO [stdout] 1 25.0    2009-07-01  2011-12-31  1   2009-07-01  2011-12-31  
INFO [stdout] 1 27.0    2012-01-01  2030-12-31  1   2012-01-01  2030-12-31

现在缺少的行也发生了变化。

Answer 1

如果您注意到，每次都缺少两行。所以，首先将光标移动到第一行

joinRowSet.first();

然后，在开始实际循环之前，您已经两次调用next并且没有打印

if (!joinRowSet.next() ) {
           ...
           ...
     while(joinRowSet.next()) {

joinRowSet.next() 会将光标从当前位置移动一行。所以，当你真正开始打印你的行时，你已经到了第三行。

使用JoinRowSet时缺少两行

1 个答案: