Develop Reference

Execute a SQL query in a loop

I have this scenario. I will trigger a job in the server and as soon as the job is triggered an entry will be made into the job table with Execution_status_code as 1. I need to wait for some time say 5 mins and recheck the Execution_status_code value. As soon as the value is changed to 2, I need to proceed further.
I am using an existing connection for connecting to the database. I need to execute the SQL and if the SQL output is In progress, I need to wait for some time and then again execute the statement. Do this until the SQL output is success, until then keep waiting.
Below is the code I have tried.
Thread t = new Thread();
java.sql.Connection conn_javaComp = (java.sql.Connection)globalMap.get("conn_tNetezzaConnection_1");
java.sql.Statement st = null;
java.sql.ResultSet rs = null;
String check = null;
String dbquery_javaComp = "select case when EXECUTION_STATUS_CODE = 2 then 'Success' when EXECUTION_STATUS_CODE = 1 then 'In progress' else 'Failure' end as EXECUTION_STATUS_CODE from JOB_BKUP_NCR where JOB_TYPE_CODE="+context.JobTypeCode+" and Load_id = (select max(load_id) from JOB_BKUP_NCR where job_type_code="+context.JobTypeCode+") and START_DATETIME = (select max(START_DATETIME) from JOB_BKUP_NCR where job_type_Code="+context.JobTypeCode+")";
try
{
do
{
st = conn_javaComp.createStatement();
rs = st.executeQuery(dbquery_javaComp);
if(rs.next())
{
check = rs.getString(1);
System.out.println(check);
if (check.equalsIgnoreCase("In Progress"))
{
t.sleep(1000);
System.out.println("thread executed1");
System.out.println(dbquery_javaComp);
System.out.println(check);
}
}
else {
System.out.println(" No data found");
}
}while (!"Success".equals(check));
}
catch (Exception e) {
e.printStackTrace();
} finally {
try {
if( rs != null)
rs.close();
if( st!= null)
st.close();
}
catch (Exception e1) {
e1.printStackTrace();
}
}
The output i am getting is 'In Progress'. The loop is struck at In progress even after i change the value in the database. I am not sure where i am doing wrong. Any suggestions?

You are creating a new statement and a new resultset inside the loop, and so, they should be close inside the loop. I am thinking that your connection got corrupted with multiple statements and resultset without closing them. Please try to close them and see if that work.

The data that you are seeing will be cached.
Try closing and re-opening your DB connection. This may not even be good enough if you are using DB pooling.

There are many things I can foresee going wrong with your code. For once most DBMS will either lock the rows until you commit / close the connection or give you a snapshot of the data instead, hence you don't see the updated value or the transaction that supposed to update it wouldn't go through. Try comitting or close/reopen the transaction per loop iteration.
I would also doubt if this is a good code design as you are doing "polling". Consider if you can find other method of getting notified of the event.

try
{
//declare here your statement and resultset
st = conn_javaComp.createStatement();
rs = st.executeQuery(dbquery_javaComp);
do
{
if(rs.next())
{
check = rs.getString(1);
System.out.println(check);
if (check.equalsIgnoreCase("In Progress"))
{
t.sleep(1000);
System.out.println("thread executed1");
System.out.println(dbquery_javaComp);
System.out.println(check);
}
}
else {
System.out.println(" No data found");
}
}while (!"Success".equals(check));

Java, code works, but exception is still thrown

I am creating a simple registration frame that adds records onto a database. It gives me an error message every time it runs the SQL query that adds records in the database, however it still adds them, but because of that my programs gets to a standstill, instead of opening another window.
here's that part of the code:
regButton.addActionListener(new ActionListener() {
#Override public void actionPerformed( ActionEvent e ) {
//Execute when button is pressed
if( uNameField.getText().equals("")
|| new String(newPassField.getPassword()).equals("")
|| new String(repeatPassField.getPassword()).equals("") ) {
errorLabel.setForeground(Color.red);
errorLabel.setText("Some fields are left blank");
}
else if( new String(newPassField.getPassword()).equals(
new String(repeatPassField.getPassword()))){
Statement stmt;
ResultSet res;
try
{
//SET USERNAME AND PASSWORD FROM FIELDS TO UPPER CASE
String username = uNameField.getText().toUpperCase();
String password = new String(newPassField.getPassword()).toUpperCase();
//SQL INSERT QUERY
String sql;
sql = "INSERT INTO Employees VALUES ('" +username +"','" +password +"');";
stmt = con.createStatement();
res = stmt.executeQuery(sql);
System.out.println("Added to database!");
con.close();
}
catch(SQLException exe) {
System.out.println("Error creating or running statement: " + e.toString());
try {
con.close();
}
catch(Exception eex){}
}
}
else {
errorLabel.setForeground(Color.red);
errorLabel.setText("Password missmatch");
}
}
Every time it registers a new employee (user) it displays this "Error creating or running statement: ..... " although, I can find the newly added employees in the employee list.
What may be causing this problem?

Before we get to your specific problem, some general advice:
Connection con = ...
try {
// your stuff
}
catch (Exception e) {
e.printStackTrace();
}
finally {
try {
con.close();
} catch (Exception e) {
e.printStackTrace();
}
}
The way you are doing it now not only swallows the exception, but also avoids printing its stacktrace. And close must be performend once and only once, regardless of the exception.
If you are on Java 7, this would be much easier:
try (Connetion con = ...) {
// stuff to do
}
catch (Exception e) {
e.printStackTrace();
}
The closing in a finally is now done automatically.
Specifically about your exception, you execute an INSERT by calling executeQuery. This method sends the statement to the DB, which properly executes it, but its response back to the JDBC is not a ResultSet. This is where it blows up, after the record is already inserted. Since you are in autocommit mode, there is no transaction to roll back. Lesson: always use transactions.

You need to use executeUpdate for SQL INSERTs
int rowCount = stmt.executeUpdate(sql);

I hate seeing code written this way. You didn't ask about this, and my comment won't solve your problem, but I think it needs to be said.
You're creating a maintenance nightmare for yourself by putting persistence code in a Swing Listener method.
A better idea is to think about objects in a way that gives them a single responsibility.
Take your persistence code and move it into a separate class that you can develop and test on its own. Once it's working, give a reference to the class that needs it.
Your code will be more modular, easier to test, more reusable, and less of a nightmare to understand.
Uncle Bob Martin has a succinct mneumonic for this and other ideas worth remembering: SOLID.

why dont you try PreparedStatement
try{
//SET USERNAME AND PASSWORD FROM FIELDS TO UPPER CASE
String username = uNameField.getText().toUpperCase();
String password = new String(newPassField.getPassword()).toUpperCase();
//SQL INSERT QUERY
PreparedStatement pstmt = con.prepareStatement("insert into Employees values(?,?)");
pstmt.setString(1,username);
pstmt.setString(2,password);
if(!pstmt.execute())
{
//means your code worked correctly
System.out.println("Inserted successfully");
}
else
{
System.out.println("Unsuccessfull");
}
}
catch(Exception ex)
{
ex.printStackTrace();
}

jdbc performance

There are thee tables inside my database. One is employee, the second is employee_Project, and the third is employee_Reporting. Each table has a common employee_Number as its primary key, and there is a one to many relationship among them such that an employee has many projects and reporting dates.
I have run select * from employee, select * from employee_project, select * from employee_reporting in three data holder classes which have methods fillResultSet(Result set) and List<T> getData(). This is based on a SqlDbEngine class with a runQuery(PreparedStatement,DataHolder) method, and the implementation has been completed.
Now I have to design a getAllEmployee() method along with project and reporting detail with optimal code in java using JDBC. I have used an iterator but this solution is not acceptable; now I have to use a foreach loop.
This is what I have done:
public List<Employee> getAllEmployees() {
EmployeeDataHolderImpl empdataholder = new EmployeeDataHolderImpl();
List<Employee> list_Employee_Add = null;
try {
Connection connection = mySqlDbConnection.getConnection();
PreparedStatement preparedStatement = connection
.prepareStatement(GET_ALL_EMPLOYEE_DETAILS);
mySqlDBEngineImpl.runQuery(preparedStatement, empdataholder);
} catch (SQLException e) {
e.printStackTrace();
}
for (Employee employee : empdataholder.getData()) {
new EmployeeDAOImpl().getProject(employee);
new EmployeeDAOImpl.getReport(employee);
}
list_Employee_Add = empdataholder.getData();
return list_Employee_Add;
}
and make another method
public void getProject(Employee emp) {
EmployeeProjectDataHolderImpl employeeProjectHolder = new EmployeeProjectDataHolderImpl();
try {
Connection connection = mySqlDbConnection.getConnection();
PreparedStatement preparedStatement = connection
.prepareStatement(GET_ALL_PROJECT_DETAILS);
mySqlDBEngineImpl
.runQuery(preparedStatement, employeeProjectHolder);
} catch (SQLException e) {
e.printStackTrace();
}
for (EmployeeProject employee_Project : employeeProjectHolder.getData()) {
if (employee_Project.getEmployeeNumber() == emp.getEmpNumber()) {
emp.getProjects().add(employee_Project);
}
}
}
public void getReport(Employee emp) {
EmployeeReportDataHolderImpl employeeReportHolder = new EmployeeReportDataHolderImpl();
try {
Connection connection = mySqlDbConnection.getConnection();
PreparedStatement preparedStatement = connection
.prepareStatement(GET_ALL_REPORT_DETAILS);
mySqlDBEngineImpl
.runQuery(preparedStatement, employeeReportHolder);
} catch (SQLException e) {
e.printStackTrace();
}
for (EmployeeReport employee_Report : employeeReportHolder.getData()) {
if (employee_Report.getEmployeeNumber() == emp.getEmpNumber()) {
emp.getProjects().add(employee_Project);
}
}
}
}
and same for Employee Reporting but doing, this performance is going to decrease.no body worry about closing connection i will do it
Please tell me how I could improve my solution..

There are some issue with your code.
1.you are initializing EmployeeDAOImpl everytime, rather you can just keep one instance and call the operations over it.
new EmployeeDAOImpl().getProject(employee); new
EmployeeDAOImpl.getReport(employee);
2.I don't see where you close your connection after performing an SQL operation.

You should be having
try {
--code statements
}
catch(SQLException e){
e.printStackTrace();
}
finally{
-- close your connection and preparedStatement
}
Closing database connections is very vital.

If you use your actual code, you will have 3 impacts in your code:
You're opening a connection to get the employee's data.
For every employee, you open (and close) a new connection to get his projects.
For every employee, you open (and close) a new connection to get his reports.
Note that opening a new connection is a performance hit on your application. It doesn't matter if you use an enhanced for-loop or an Iterator, there would be many hits that can slow down your application.
Two ways to solve this problem:
Open a single connection where you run all your select statements. This will be better than opening/closing lot of connections.
Create a single SQL statement to retrieve the employees and the data you need for every employee. It will have better performance for different reasons:
A single connection to the database.
A single query instead of lot of queries to the database (a single I/O operation).
If your rdbms allows it, the query will be optimized for future requests (a single query instead of multiple queries).
I would prefer to go with the second option. For this, I tend to use a method that executes any SQL select statement and return a ResultSet. I'll post a basic example (note, the provided code can be improved depending on your needs), this method could be in your SqlDbEngine class:
public ResultSet executeSQL(Connection con, String sql, List<Object> arguments) {
PreparedStatement pstmt = null;
ResultSet rs = null;
try {
pstmt = con.prepareStatement(sql);
if (arguments != null) {
int i = 1;
for(Object o : arguments) {
pstmt.setObject(i++, o);
}
}
//method to execute insert, update, delete statements...
rs = pstmt.execute();
} catch(SQLException e) {
//handle the error...
}
return rs;
}
And this other method to handle all the query operation
public List<Employee> getAllEmployee() {
Connection con = null;
ResultSet rs = null;
List<Employee> lstEmployee = new ArrayList<Employee>();
try {
con = mySqlDbConnection.getConnection();
//write the sql to retrieve all the data
//I'm assuming these can be your columns, it's up to you
//this can be written using JOINs...
String sql = "SELECT E.EMPLOYEE_ID, E.EMPLOYEE_NAME, P.PROJECT_NAME, R.REPORT_NAME FROM EMPLOYEE E, PROJECT P, REPORT R WHERE E.EMPLOYEE_ID = P.EMPLOYEE_ID AND E.EMPLOYEE_ID = R.EMPLOYEE_ID";
//I guess you don't need parameters for this...
rs = SqlDbEngine.executeSQL(con, sql, null);
if (rs != null) {
Employee e;
int employeeId = -1, lastEmployeeId = -1;
while (rs.next()) {
//you need to make sure to create a new employee only when
//reading a new employee id
employeeId = rs.getInt("EMPLOYEE_ID");
if (lastEmployeeId != employeeId) {
e = new Employee();
lastEmployeeId = employeeId;
lstEmployee.add(e);
}
Project p = new Project();
Report r = new Report();
//fill values of p...
//fill values of r...
//you can fill the values taking advantage of the column name in the resultset
//at last, link the project and report to the employee
e.getProjects().add(p);
e.getReports().add(r);
}
}
} catch (Exception e) {
//handle the error...
} finally {
try {
if (rs != null) {
Statement stmt = rs.getStatement();
rs.close();
stmt.close();
}
if (con != null) {
con.close();
}
} catch (SQLException e) {
//handle the error...
}
}
return lstEmployee;
}
Note that the second way can be harder to code but it will give you the best performance. It's up to you to improve the provided methods, some advices:
Create a class that receives a ResultSet and builds a Project instance using the columns name of the ResultSet (similar for Report and Employee).
Create a method that handles the ResultSet and its Statement close.
As a best practice, never use select * from mytable, it's preferable to write the needed columns.

If I understand correctly, your code first loads all EmployeeReport rows and then filters them according to getEmployeeNumber(). You can let your database do this by modifying your SQL query.
Since you didn't show your SQL queries (I assume they're in GET_ALL_REPORT_DETAILS), I'll just make a guess... Try executing SQL like:
select *
from employee_reporting
where employeeNumber = ?
If you put this in a PreparedStatement, and then set the parameter value, your database will only return the data you need. For example:
PreparedStatement pstmt = con.prepareStatement(GET_ALL_REPORT_DETAILS);
pstmt.setInt(1, employee.getEmployeeNumber());
That should return only the EmployeeReport records having the desired employeeNumber. In case performance is still an issue, you could consider adding an index to your EmployeeReport table, but that's a different story...

problems with update statement in SQLite

I have created a database using SQLite. I want to update the value of a "features" column( type Blob)...but i do not know how to write the "update" statement .
This is what i tried:
try {
stat = conn.createStatement();
} catch (SQLException e) {
}
try {
byte[] b = getFunction();
stat.executeUpdate("update table set features="+b);
} catch (SQLException e) {
}
i get the follwing error :
java.sql.SQLException: unrecognized token: "[B#13a317a"
so i guess that "b" is the problem ?

[B#13a317a looks like a array to string result (b.toString() in this case). You should use a prepared statement for the blob like:
update table set features=?
An example is here.
Generally, you should never create a SQL by concatenating strings. This is the recipe for SQL injection problems.

Try this one with PreparedStatement:
Connection con = null;
PreparedStatement stmt = null;
try {
byte[] b = getFunction();
con = ...;
stmt = con.prepareStatement("update table set features=?");
stmt.setBytes(1, b);
stmt.executeUpdate();
con.commit();
}
catch (SQLException e) {
//handle exception (consider con.rollback()) and con maybe null here)
}
finally {
//close stmt and at least con here (all maybe null here)
}
Personally I am always using PreparedStatements. When you have to write a lot of this code then consider writing some Utility-Classes to reduce Boilerplate-Code.
In particular you should consider writing Utilty-Classes for null-safe calling methods on Connection, Statement and ResultSet methods when you are dealing with plain JDBC.
EDIT
What Thomas Jung wrote about preventing SQL Injections is another big pro for always using PreparedStatements. +1 for him :-)

stat.executeUpdate("update table set features="+b[0].toString());
you have to use +

exhausted resultset error when i get the name of the columns

i'm trying to get the type and the name of the result and when enter in the loop, excuting somo instructions about the metadata the resulset.next changed from true to false, and give the error java.sql.SqlExcepcion exhausted resultset. Any ideas? i really dont know how solved it because i read the post with the solution of this problem and validate if the resultset it's null before begin the loop. I'm called this method with a scheduler of quartz. I'm using this in a j2ee aplication and the example it's this
try
{
InitialContext ctx = new InitialContext();
WrapperDataSource wrapperDataSource = (WrapperDataSource)ctx.lookup(systemLogger.getConfigurationParameters().getDataSource());
conn = wrapperDataSource.getConnection();
Class.forName("com.microsoft.sqlserver.jdbc.SQLServerDriver").newInstance();
conn = DriverManager.getConnection(url,login,password);
if (conn != null)
{
stmt = conn.createStatement();
res = stmt.executeQuery(query);
if (res != null)
{
while (res.next())
{
for (int i = 0; i < columnlength; i++)
{
String columnName = metadata.getColumnName(i+1);
if (metadata.getColumnName(i+1).equalsIgnoreCase(systemLogger.getColumnStatus()))
{
columnStatusType = metadata.getColumnType(i+1);
}
else if (metadata.getColumnName(i+1).equalsIgnoreCase(systemLogger.getColumnDocumentId()))
{
columnDocumentIdType = metadata.getColumnType(i+1);
}
else if (metadata.getColumnName(i+1).equalsIgnoreCase(systemLogger.getColumnTimer()))
{
columnTimerType = metadata.getColumnType(i+1);
}
}
}
}
else
{
__log.error("No results found for the query");
throw new PtmServiceException("No se encontraron resultados para el query");
}
}
else
{
__log.error("Could not create the connection");
throw new PtmServiceException("No se pudo crear la conexion");
}
}
catch(Exception e)
{
__log.error("Error in the execution of the query");
throw new PtmServiceException("Error ejecutando la busqueda");
}
finally
{
res.close();
stmt.close();
conn.close();
}

The variable columnlength seems to hold a value larger than the number of columns returned by the query. Try with a smaller columnlength.

finally, i see the problem, while i'm debugging the code with ecplise in the view of the expressions i added the follow expression res.next(), then each sentence that i pass for the step into bring the consequence that expression that evaluate if the resultset has more rows, be evaluated again. In some point the resultset has evaluated all the rows for each step into that i made in the process of debugging. The only thing that i have to do was eliminate the expression and works fine...

The problem might not be with the code but instead could be the database. Double check that the TABLE IS NOT EMPTY. You get this error if the table is empty. Keep in mind that databases like Oracle require a commit after all your insert, update, alter statements .Your changes might not be visible outside the database till you run a commit over the your db, I was having this problem for quite a long time. I kept on checking the table with select statement but the problem with my oracle db was that I had not issued a commit over my db.