I have to modify a few tables in one function. They must all succeed, or all fail. If one operation fails, I want them all to fail. I have the following:
public void foo() throws Exception {
Connection conn = null;
try {
conn = ...;
conn.setAutoCommit(false);
grok(conn);
conn.commit();
}
catch (Exception ex) {
// do I need to call conn.rollback() here?
}
finally {
if (conn != null) {
conn.close();
conn = null;
}
}
}
private void grok(Connection conn) throws Exception {
PreparedStatement stmt = null;
try {
// modify table "apple"
stmt = conn.prepareStatement(...);
stmt.executeUpdate();
stmt.close();
// modify table "orange"
stmt = conn.prepareStatement(...);
stmt.executeUpdate();
stmt.close();
...
}
finally {
if (stmt != null) {
stmt.close();
}
}
}
I'm wondering if I need to call rollback() in the case that something goes wrong during this process.
Other info: I'm using connection pooling. In the sample above, I'm also making sure to close each PreparedStatement using finally statements as well, just left out for brevity.
Thank you
Yes you need to call rollback if any of your statements fails or you have detected an exception prior to calling commit. This is an old post but the accepted answer is wrong. You can try it for yourself by throwing an exception before commit and observing that your inserts still make it into the database if you do not manually rollback.
JDBC Documentation
https://docs.oracle.com/javase/tutorial/jdbc/basics/transactions.html#call_rollback
Example Correct Usage from the doc
public void updateCoffeeSales(HashMap<String, Integer> salesForWeek)
throws SQLException {
PreparedStatement updateSales = null;
PreparedStatement updateTotal = null;
String updateString =
"update " + dbName + ".COFFEES " +
"set SALES = ? where COF_NAME = ?";
String updateStatement =
"update " + dbName + ".COFFEES " +
"set TOTAL = TOTAL + ? " +
"where COF_NAME = ?";
try {
con.setAutoCommit(false);
updateSales = con.prepareStatement(updateString);
updateTotal = con.prepareStatement(updateStatement);
for (Map.Entry<String, Integer> e : salesForWeek.entrySet()) {
updateSales.setInt(1, e.getValue().intValue());
updateSales.setString(2, e.getKey());
updateSales.executeUpdate();
updateTotal.setInt(1, e.getValue().intValue());
updateTotal.setString(2, e.getKey());
updateTotal.executeUpdate();
con.commit();
}
} catch (SQLException e ) {
JDBCTutorialUtilities.printSQLException(e);
if (con != null) {
try {
System.err.print("Transaction is being rolled back");
con.rollback();
} catch(SQLException excep) {
JDBCTutorialUtilities.printSQLException(excep);
}
}
} finally {
if (updateSales != null) {
updateSales.close();
}
if (updateTotal != null) {
updateTotal.close();
}
con.setAutoCommit(true);
}
}
You don't need to call rollback(). If the connection closes without completing commit() it will be rolled back.
You don't need to set conn to null either; and since the try block starts after conn is initialized (assuming ... cannot evaluate to null) you don't need the != null in finally either.
If you call "commit" then the transaction will be committed. If you have multiple insert/update statements and one of them fails, committing will cause the inserts/updates that didn't fail to commit to the database. So yes, if you don't want the other statements to commit to the db, you need to call rollback. What you are essentially doing by setting autocommit to false is allowing multiple statements to be committed or rolledback together. Otherwise each individual statement will automatically commit.
Related
If you running a JDBC program from your laptop which is connected to office network over wi-fi.
What happens if network is lost while looping through ( JDBC ) result set?
I noticed java.sql.Connection & Statement objects have timeout. But result set has no time out setting.
ResultSet rs = st.getResultSet();
while (rs.next()) {
int id = rs.getInt(1);
// use this and run some biz logic.
}
I noticed program keep waiting for next results forever. how to make it throw exception ? Has any one experienced this? What did you do ?
ResultSet will not be longer usable, but what you can do is:
You have two options:
First: (and I suggest) using Transactions, As Oracle doc says:
When a connection is created, it is in auto-commit mode. This means
that each individual SQL statement is treated as a transaction and is
automatically committed right after it is executed. (To be more
precise, the default is for a SQL statement to be committed when it is
completed, not when it is executed. A statement is completed when all
of its result sets and update counts have been retrieved. In almost
all cases, however, a statement is completed, and therefore committed,
right after it is executed.)
And in your case you should execute once after all of your statement are completed, in case of network failure or any error in one statement all of your statements will rolled back
Example: reference
public void updateCoffeeSales(HashMap<String, Integer> salesForWeek)
throws SQLException {
PreparedStatement updateSales = null;
PreparedStatement updateTotal = null;
String updateString =
"update " + dbName + ".COFFEES " +
"set SALES = ? where COF_NAME = ?";
String updateStatement =
"update " + dbName + ".COFFEES " +
"set TOTAL = TOTAL + ? " +
"where COF_NAME = ?";
try {
con.setAutoCommit(false);
updateSales = con.prepareStatement(updateString);
updateTotal = con.prepareStatement(updateStatement);
for (Map.Entry<String, Integer> e : salesForWeek.entrySet()) {
updateSales.setInt(1, e.getValue().intValue());
updateSales.setString(2, e.getKey());
updateSales.executeUpdate();
updateTotal.setInt(1, e.getValue().intValue());
updateTotal.setString(2, e.getKey());
updateTotal.executeUpdate();
con.commit();
}
} catch (SQLException e ) {
JDBCTutorialUtilities.printSQLException(e);
if (con != null) {
try {
System.err.print("Transaction is being rolled back");
con.rollback();
} catch(SQLException excep) {
JDBCTutorialUtilities.printSQLException(excep);
}
}
} finally {
if (updateSales != null) {
updateSales.close();
}
if (updateTotal != null) {
updateTotal.close();
}
con.setAutoCommit(true);
}
}
Second: You may have to buffer all the results before the connection was lost but take care of closing the result set, statement and connection
Im doing a Oracle Select for Update with java and it works on times and sometimes it hangs with the session and cannot remove the locked session (have to manually kill the session )
this works fine for most of the scenarios but when I deployed it in two servers ( web service ) and request them both at once this happens , I can't understand whether it's a problem with my code ,
my code
public boolean checkJobStatus(long taskId)
{
Connection con = null;
PreparedStatement selectForUpdate = null;
String lastJobStatus = null;
boolean runNow = false;
try
{
con = conPool.getConnection();
con.setAutoCommit(false);
selectForUpdate = con.prepareStatement("SELECT LAST_JOB_STATUS FROM ADM_JOB WHERE TASK_ID = ? FOR UPDATE ");
selectForUpdate.setLong(1, taskId);
ResultSet resultSet = selectForUpdate.executeQuery();
while(resultSet.next())
{
if (resultSet.getObject("LAST_JOB_STATUS") == null)
{
lastJobStatus = ScheduledJob.STATUS_FAILED;
}
else
{
lastJobStatus = resultSet.getString("LAST_JOB_STATUS");
}
}
if(ScheduledJob.STATUS_RUNNING.equalsIgnoreCase(lastJobStatus) || ScheduledJob.STATUS_STARTED.equalsIgnoreCase(lastJobStatus))
{
runNow = false;
// commit n update setting autocommit to true
selectForUpdate = con.prepareStatement("UPDATE ADM_JOB SET LAST_JOB_STATUS =? WHERE TASK_ID = ?");
selectForUpdate.setString(1, lastJobStatus);
selectForUpdate.setLong(2, taskId);
selectForUpdate.executeUpdate();
}
else
{
runNow =true;
// commit n update setting autocommit to true
selectForUpdate = con.prepareStatement("UPDATE ADM_JOB SET LAST_JOB_STATUS =? WHERE TASK_ID = ?");
selectForUpdate.setString(1, ScheduledJob.STATUS_STARTED);
selectForUpdate.setLong(2, taskId);
selectForUpdate.executeUpdate();
con.commit();
con.setAutoCommit(true);
}
} catch (SQLException e)
{
Logger.getLogger( "" ).log(Level.SEVERE, "Error in getting database connection", e);
try
{
con.rollback(); // rolling back the row lock in case of a exception
} catch (SQLException e1)
{
e1.printStackTrace();
}
}
finally
{
DBUtility.close( selectForUpdate );
DBUtility.close( con );
}
return runNow;
}
Commit occurs only in the else branch. If this condition doesn't happen, transaction is not closed, so a second thread hangs up forever on the select for update.
I'm trying to get information from a MySQL database. I can connect and do things such as insert data into tables fine, and although I receive a ResultSet, I can't read it. Here's my code:
public ResultSet executeQuery(String query) {
Statement stmt = null;
ResultSet rs = null;
try {
stmt = conn.createStatement();
rs = stmt.executeQuery(query);
if (stmt.execute(query)) {
rs = stmt.getResultSet();
}
return rs;
}
catch (SQLException ex){
System.err.println("SQLException: " + ex.getMessage());
System.err.println("SQLState: " + ex.getSQLState());
System.err.println("VendorError: " + ex.getErrorCode());
}
finally {
if (stmt != null) {
try {
stmt.close();
} catch (SQLException sqlEx) { } // ignore
stmt = null;
}
}
return null;
}
Trying to read the ResultSet:
ResultSet set = executeQuery("SELECT rank FROM players");
try {
while(set.next()) {
System.out.println(set.getInt("rank") + "");
}
} catch(Exception ex) {
ex.printStackTrace();
} finally {
try {
set.close();
} catch (SQLException e1) {
e1.printStackTrace();
}
}
I get this error:
java.sql.SQLException: Operation not allowed after ResultSet closed
I've been looking around the internet and on different forums all day. What's wrong with my code?
Your finally block in executeQuery closes your statement before you've iterated over the results.
Well, two things -
1) I do not get why you have both execute() and executeQuery() in your method? I think executeQuery() should work for you.
2) Do not close the statement before iterating through resultset. Make it an instance variable and close it after iterating.
TL;DR Your executeQuery closes the Statement.
From the JavaDoc for Statement.close
Releases this Statement object's database and JDBC resources
immediately instead of waiting for this to happen when it is
automatically closed.
You close() your Statement before you read your ResultSet. Closing a Statement causes all of the ResultSet instances associate with that Statement to also be closed.
You need to pass in a Consumer<ResultSet> into a different method, and process the ResultSet whilst the Statement is open:
public void executeQuery(final String query, final Consumer<ResultSet> consumer) {
try(final Statement stmt = conn.createStatement();
final ResultSet rs = stmt.executeQuery(query)) {
consumer.accept(rs);
}
catch (SQLException ex){
System.err.println("SQLException: " + ex.getMessage());
System.err.println("SQLState: " + ex.getSQLState());
System.err.println("VendorError: " + ex.getErrorCode());
}
}
I have also fixed your code:
I have replaced your try..finally with a try-with-resources
You execute your query twice by calling stmt.executeQuery(query) and then stmt.execute(query). This is not only wasteful, but you lose the reference to the first ResultSet and if Statement.close didn't close all the associated resultsets you would have had a memory leak.
I am developing high-load application using tomcat jdbc connection pool and Oracle database. It is very important to ensure my app to have very small DB query timeouts (no longer than 3 seconds) to prevent long-running queries or database slowness from blocking all my application. To simulate long-running queries I have put the DB in QUIESCE state using ALTER SYSTEM QUIESCE RESTRICTED statement.
But it looks like the timeout values have no impact - when i begin to test my application, it hangs...
Here is my jdbc pool configuration:
String connprops = "oracle.net.CONNECT_TIMEOUT=3000;oracle.jdbc.ReadTimeout=3000;"
+ "oracle.net.READ_TIMEOUT=3000";
pp.setConnectionProperties(connprops);
pp.setDriverClassName("oracle.jdbc.OracleDriver");
pp.setTestOnBorrow(true);
pp.setTestOnConnect(true);
pp.setTestOnReturn(true);
pp.setTestWhileIdle(true);
pp.setMaxWait(2000);
pp.setMinEvictableIdleTimeMillis(20000);
pp.setTimeBetweenEvictionRunsMillis(20000);
pp.setValidationInterval(3000);
pp.setValidationQuery("SELECT 1 FROM DUAL");
pp.setMaxAge(3000);
pp.setRemoveAbandoned(true);
pp.setRemoveAbandonedTimeout(3);
pp.setJdbcInterceptors("org.apache.tomcat.jdbc.pool.interceptor.QueryTimeoutInterceptor(queryTimeout=3)");
dataSource = new DataSource();
dataSource.setPoolProperties(pp);
That's how i work with connections (pretty simple):
Connection conn = dataSource.getConnection();
Statement stmt = null;
ResultSet rs = null;
try {
stmt = conn.createStatement();
rs = stmt.executeQuery(/*some select query*/);
if (rs.next()) {
result = rs.getInt(1);
/*process the result*/
}
rs.close();
stmt.close();
conn.close();
}
catch(Exception e) {
logger.error("Exception: " + e.getMessage(), e);
}finally {
if (conn != null) {
if(rs!=null)
rs.close();
if(stmt!=null)
stmt.close();
conn.close();
}
}
Any ideas? Thanks in advance!
Try to use this config:
String connprops = "oracle.net.CONNECT_TIMEOUT=\"3000\";oracle.jdbc.ReadTimeout=\"3000\";"
+ "oracle.net.READ_TIMEOUT=\"3000\"";
All non-string values are ignored by java.util.Properties.java:
public String getProperty(String key) {
Object oval = super.get(key);
String sval = (oval instanceof String) ? (String)oval : null; // <- !!!!
return ((sval == null) && (defaults != null)) ? defaults.getProperty(key) : sval;
}
You should probably also use java.sql.Statement's query timeout:
stmt.setQueryTimeout(3); // int seconds
This question already has answers here:
How should I use try-with-resources with JDBC?
(5 answers)
Closed 8 years ago.
Yesterday multiple people on Stack recommended using try-with-resources. I am doing this for all my database operations now. Today I wanted to change Statement to PreparedStatement to make the queries more secure. But when I try to use a prepared statement in try-with-resources I keep getting errors like 'identifier expected' or ';' or ')'.
What am I doing wrong? Or isnt this possible? This is my code:
try (Connection conn = DriverManager.getConnection(DBURL, DBUSER, DBPASS);
PreparedStatement stmt = conn.prepareStatement("SELECT id FROM users WHERE id = ? LIMIT 1");
stmt.setInt(1, user);
ResultSet rs = stmt.executeQuery()) {
// if no record found
if(!rs.isBeforeFirst()) {
return false;
}
// if record found
else {
return true;
}
} catch (SQLException e) {
// log error but dont do anything, maybe later
String error = "SQLException: " + e.getMessage() + "\nSQLState: " + e.getSQLState() + "\nVendorError: " + e.getErrorCode();
return false;
}
A try-with-resource statement is used to declare (Autoclosable) resources. Connection, PreparedStatement and ResultSet are Autoclosable, so that's fine.
But stmt.setInt(1, user) is NOT a resource, but a simple statement. You cannot have simple statements (that are no resource declarations) within a try-with-resource statement!
Solution: Create multiple try-with-resource statements!
try (Connection conn = DriverManager.getConnection(DBURL, DBUSER, DBPASS)) {
executeStatement(conn);
} catch (SQLException e) {
// log error but dont do anything, maybe later
String error = "SQLException: " + e.getMessage() + "\nSQLState: " + e.getSQLState() + "\nVendorError: " + e.getErrorCode();
return false;
}
private void executeStatement(Connection con) throws SQLException {
try (PreparedStatement stmt = conn.prepareStatement("SELECT id FROM users WHERE id=? LIMIT 1")) {
stmt.setInt(1, user);
try (ResultSet rs = stmt.executeQuery()) {
// process result
}
}
}
(Please note that technically it is not required to put the execution of the SQL statement into a separate method as I did. It also works if both, opening the connection and creating the PreparedStatement are within the same try-with-resource statement. I just consider it good practice to separate connection management stuff from the rest of the code).
try this code:
try (Connection conn = DriverManager.getConnection(DBURL, DBUSER, DBPASS)) {
PreparedStatement stmt = conn.prepareStatement("SELECT id FROM users WHERE id = ? LIMIT 1");
stmt.setInt(1, user);
ResultSet rs = pstmt.executeQuery())
// if no record found
if(!rs.isBeforeFirst()) {
return false;
}
// if record found
else {
return true;
}
} catch (SQLException e) {
// log error but dont do anything, maybe later
String error = "SQLException: " + e.getMessage() + "\nSQLState: " + e.getSQLState() + "\nVendorError: " + e.getErrorCode();
return false;
}
note that here, resource is your Connection and you have to use it in the try block ()
Move
stmt.setInt(1, user);
ResultSet rs = stmt.executeQuery()
...within the try{ /*HERE*/ }
This is because stmt is the resource being created try (/*HERE*/) {} to be used try{ /*HERE*/ }
Try-with-resources
try (/*Create resources in here such as conn and stmt*/)
{
//Use the resources created above such as stmt
}
The point being that everything created in the resource creation block implements AutoClosable and when the try block is exited, close() is called on them all.
In your code stmt.setInt(1, user); is not an AutoCloseable resource, hence the problem.