JDBC如何使用多個值查詢從單個插入中獲取所有生成的密鑰？

Question

我使用PreparedStatement使用單個查詢中插入多行：

String query = "insert into MyTable (a,b,c) values (?,?,?),(?,?,?),(?,?,?),(?,?,?)"; // insert 4 rows in a single query 
PreparedStatement stmt = connection.prepareStatement(query, Statement.RETURN_GENERATED_KEYS); 
// .. here: loop to set all 4 x 3 values in my prepared statement 

而且我想在MSSQL數據庫生成的ID。我不確定execute使用哪種味道。

的​​方法不返回預期ResultSet：

boolean thingy = stmt.execute(); 
System.out.println("execute returned ", thingy); // false: meaning no result set 
ResultSet result = stmt.getResultSet(); // result is null 

的executeUpdate()方法只返回ResultSet：

int rowCount = stmt.executeUpdate(); 
System.out.println(rowCount + " rows updated"); // 4 rows updated 
ResultSet result = stmt.getGeneratedKeys(); 
// there's only one result: 
result.next(); // true 
int ID1 = result.getInt(1); // good 
result.next(); // false, no more rows 

是否有可能與一個單一的INSERT得到生成的ID？出於性能原因，我不想發送多個INSERT查詢。

DB：MS SQL服務器11（2012）
網址爲：jdbc：JTDS：SQLSERVER：// ...
司機：net.sourceforge.jtds.jdbc.Driver JTDS版本1.3.1

Answer 1

在寫這個答案的時候，不JTDS也不是MSSQL-JDBC支持返回多個生成使用Statement.RETURN_GENERATED_KEYS鍵，無論是從使用TVC（表值構造函數）INSERT語句或使用executeBatch用PreparedStatement。但是，下面的代碼可以使用兩種驅動程序（在寫作的時間JTDS 1.3.1和the dev branch of mssql-jdbc測試）：

try (Statement st = conn.createStatement()) { 
    // TEST ENVIRONMENT: in a real application, this would be our permanent table 
    st.execute("CREATE TABLE #tmp (id INT IDENTITY PRIMARY KEY, txtcol NVARCHAR(50))"); 
    st.execute("INSERT INTO #tmp (txtcol) VALUES (N'existing data')"); 
    st.execute("INSERT INTO #tmp (txtcol) VALUES (N'more existing data')"); 
} 
StringBuilder sb = new StringBuilder(); 
sb.append("SET NOCOUNT ON; "); 
sb.append("DECLARE @seq INT, @txt NVARCHAR(50), @id INT; "); 
sb.append("DECLARE @newIds TABLE (seq INT, id INT); "); 
sb.append("DECLARE @toInsert TABLE (seq INT IDENTITY PRIMARY KEY, txt NVARCHAR(50)); "); 
sb.append(""); 
sb.append("INSERT INTO @toInsert (txt) VALUES (?), (?), (?); "); 
sb.append(""); 
sb.append("DECLARE crsr CURSOR FOR SELECT seq, txt FROM @toInsert ORDER BY seq; "); 
sb.append("OPEN crsr; "); 
sb.append("FETCH NEXT FROM crsr INTO @seq, @txt; "); 
sb.append("WHILE @@FETCH_STATUS = 0 "); 
sb.append("BEGIN "); 
sb.append(" INSERT INTO #tmp (txtcol) VALUES (@txt); "); 
sb.append(" SELECT @id = SCOPE_IDENTITY(); "); 
sb.append(" INSERT INTO @newIds (seq, id) VALUES (@seq, @id); "); 
sb.append(" FETCH NEXT FROM crsr INTO @seq, @txt; "); 
sb.append("END "); 
sb.append("SELECT id FROM @newIds ORDER BY seq; "); 
try (PreparedStatement ps = conn.prepareStatement(sb.toString())) { 
    ps.setString(1, "foo"); 
    ps.setString(2, "bar"); 
    ps.setString(3, "baz"); 
    try (ResultSet rs = ps.executeQuery()) { 
     while (rs.next()) { 
      System.out.println(rs.getInt(1)); 
     } 
    } 
} 

產生

3 
4 
5 

Answer 2

對於SQL Server 2008及更高版本，添加OUTPUT clause的INSERT語句，似乎這樣的伎倆：

try (Statement st = conn.createStatement()) { 
    // TEST ENVIRONMENT: in a real application, this would be our permanent table 
    st.execute("CREATE TABLE #tmp (id INT IDENTITY PRIMARY KEY, txtcol NVARCHAR(50))"); 
    st.execute("INSERT INTO #tmp (txtcol) VALUES (N'existing data')"); 
    st.execute("INSERT INTO #tmp (txtcol) VALUES (N'more existing data')"); 
} 
StringBuilder sb = new StringBuilder(); 
sb.append("INSERT INTO #tmp (txtcol) "); 
sb.append(" OUTPUT INSERTED.id "); 
sb.append(" VALUES (?), (?), (?) "); 
try (PreparedStatement ps = conn.prepareStatement(sb.toString())) { 
    ps.setString(1, "foo"); 
    ps.setString(2, "bar"); 
    ps.setString(3, "baz"); 
    try (ResultSet rs = ps.executeQuery()) { 
     while (rs.next()) { 
      System.out.println(rs.getInt(1)); 
     } 
    } 
} 

產生

3 
4 
5