过去 2-3 天我遇到了一个问题。这似乎是一个小问题，但我还没能捕获它。
问题出在代码中的结果集 rs 上。 方法 mergeTable 中的 while (rs.next()) 语句行为突然。第一次，在某些情况下它会进入 while 循环，有时不会，并且在少数情况下，当它进入 while 循环时，它会突然抛出结果集耗尽异常。我用谷歌搜索，发现结果集可能已被访问代码的其他线程关闭。但这是一个简单的独立 java 应用程序，不是多线程的。
我还注意其他两个结果集，check 和 checktarget 不会干扰结果集 rs。我也将结束报表和结果集。 你能看一下代码，看看我是否遗漏了一些东西。
使用的两个数据库实例是10.180.22.93:1521:V3demo(称为源数据库)和10.180.22.93:1521:fusiondb(称为目标数据库)。 rs 结果集来自源数据库。check 和 checktarget 结果集来自目标数据库。因此，结果集 rs 将来自源数据库的表 A，结果集 check 和 checktarget 将来自目标数据库的表 A。
static String mergeTable() throws Exception {
String result = "ERROR";
int error = 0;
String tableString = "<table " + tablename + ">";
PreparedStatement preparedSelect = null;
PreparedStatement preparedSelectTarget = null;
Statement selectSourceStmt = null;
ResultSet checkTarget = null;
ResultSet rs = null;
try {
logger.println("====================================================================================");
logger.println("Processing table:" + tablename);
System.out.println("====================================================================================");
System.out.println("Processing table:" + tablename);
// Create query to fetch records from the source
String sourceQuery = "SELECT * FROM " + tablename;
if (owner.trim().equals("F1") || owner.trim().equals("C1") || owner.trim().equals("CM"))
sourceQuery = sourceQuery + " WHERE OWNER_FLG='" + owner + "'";
// Get the result set
selectSourceStmt = source.createStatement();
rs = selectSourceStmt.executeQuery(sourceQuery);
System.out.println(sourceQuery);
String selectSQL = "SELECT COUNT(*) FROM " + tablename + " WHERE ";
String selectSQLTarget = "SELECT * FROM " + tablename + " WHERE "; // ankush
ResultSetMetaData metaData = rs.getMetaData();
List list = new ArrayList();
List typesList = new ArrayList();
for (int i = 1; i <= metaData.getColumnCount(); i++) {
String columnName = metaData.getColumnName(i);
list.add(columnName); // list contains the entire list of columns of the source
typesList.add(metaData.getColumnType(i));
}
for (int i = 1; i < keys.length; i++) {
if (i == 1) {
selectSQL = selectSQL + " " + keys[i] + "= ?";
selectSQLTarget = selectSQLTarget + " " + keys[i] + "= ?"; //ankush
}
else {
selectSQL = selectSQL + " AND " + keys[i] + "= ?";
selectSQLTarget = selectSQLTarget + " AND " + keys[i] + "= ?"; //ankush
}
}
logger.println("Select SQL:" + selectSQL);
logger.println("selectSQLTarget:" + selectSQLTarget); //ankush
preparedSelect = target.prepareStatement(selectSQL);
preparedSelectTarget = target.prepareStatement(selectSQLTarget); //ankush
int updateCount = 0, insertCount = 0, errorCount = 0;
// rs contains the entire table snapshot of source based on the owner flag
if (rs != null) {
while (rs.next()) {
try {
int i, count;
// check if record exists or not; keys contain the values of primary columns specified in the.lst file
for (int j = 1; j < keys.length; j++) {
preparedSelect.setObject(j, rs.getObject(keys[j])); // for every single row in source, corresponding rows are fetched from target.Here, where clause is being prepared
}
ResultSet check = preparedSelect.executeQuery(); // check is the target resultset for the primary key values in current row of source resultset
check.next();
count = check.getInt(1); // count gives the row/s fetched from target based on the values in source.
check.close();
// check if record exists or not; keys contain the values of primary columns specified in the.lst file
for (int j = 1; j < keys.length; j++) {
// for every single row in source, corresponding rows are fetched from target.Here, where clause is being prepared
preparedSelectTarget.setObject(j, rs.getObject(keys[j]));
}
// check is the target resultset for the primary key values in current row of source resultset
checkTarget = preparedSelectTarget.executeQuery();
checkTarget.next();
// if record exists UPDATE CONDITION
if (true) { // if there is a record in target for a row in source, update target
String rowString = "<row>";
String rowDiffFlag = "N";
// if merge flag is Y
if (mergeFlag.equals("Y")) {
String colDiffFlag = "";
String sourceColVal = "";
String targetColVal = "";
// list contains the column names
for (i = 0; i < list.size(); i++) {
System.out.println("value of i " + i);
}
i++; // ?????
} else {
logger.print("Did not update Record:");
}
rowString = rowString + "</row>";
if (rowDiffFlag.equals("Y")) {
tableString = tableString + rowString;
}
} else { // if there is no record in target for a row in source, insert into target
String sourceColVal = "";
String rowString = "<row>";
for (i = 0; i < list.size(); i++) { //looping through columns in a row
System.out.println("column " + i);
}
rowString = rowString + "</row>";
tableString = tableString + rowString;
}
} catch (Exception e1) {
e1.printStackTrace(logger);
}
}
}
} catch (Exception e) {
e.printStackTrace(logger);
} finally {
preparedSelect.close();
preparedSelectTarget.close();
selectSourceStmt.close();
checkTarget.close();
rs.close();
}
tableString = tableString + "</table>";
formXmlString(tableString);
if (error == 0) result = "SUCCESS";
return result;
}
最佳答案
哦，天哪，你这里发生的事情太多了。我看到 HTML 和 JDBC 代码混合在一起。这是个坏主意。
数据库连接不是线程安全的。您说您正在小心确保结果集不会干扰，但此错误表明情况并非如此。我会重构这段代码来隔离持久性并使一切变得更简单。这对我来说看起来是错误的，但我不愿意深入研究它来找出原因。
关于java - java 中的突然结果集行为，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/6263623/