Before we dive into deep concepts of exception handling best practices, lets start with one of the most important concepts which is to understand that there are three general types of throwable classes in Java: checked exceptions, unchecked exceptions, and errors.
1) Never swallow the exception in catch block
catch (NoSuchMethodException e) {
return null;
}
Doing this not only return “null” instead of handling or re-throwing the exception, it totally swallows the exception, losing the cause of error forever. And when you don’t know the reason of failure, how you would prevent it in future? Never do this !!
2) Declare the specific checked exceptions that your method can throw
public void foo() throws Exception { //Incorrect way
}
Always avoid doing this as in above code sample. It simply defeats the whole purpose of having checked exception. Declare the specific checked exceptions that your method can throw. If there are just too many such checked exceptions, you should probably wrap them in your own exception and add information to in exception message. You can also consider code refactoring also if possible.
public void foo() throws SpecificException1, SpecificException2 { //Correct way
}
3) Do not catch the Exception class rather catch specific sub classes
try {
someMethod();
} catch (Exception e) {
LOGGER.error("method has failed", e);
}
The problem with catching Exception is that if the method you are calling later adds a new checked exception to its method signature, the developer’s intent is that you should handle the specific new exception. If your code just catches Exception (or Throwable), you’ll never know about the change and the fact that your code is now wrong and might break at any point of time in runtime.
4) Never catch Throwable class
Well, its one step more serious trouble. Because java errors are also subclasses of the Throwable. Errors are irreversible conditions that can not be handled by JVM itself. And for some JVM implementations, JVM might not actually even invoke your catch clause on an Error.
5) Always correctly wrap the exceptions in custom exceptions so that stack trace is not lost
catch (NoSuchMethodException e) {
throw new MyServiceException("Some information: " + e.getMessage()); //Incorrect way
}
This destroys the stack trace of the original exception, and is always wrong. The correct way of doing this is:
catch (NoSuchMethodException e) {
throw new MyServiceException("Some information: " , e); //Correct way
}
6) Either log the exception or throw it but never do the both
catch (NoSuchMethodException e) {
LOGGER.error("Some information", e);
throw e;
}
As in above example code, logging and throwing will result in multiple log messages in log files, for a single problem in the code, and makes life hell for the engineer who is trying to dig through the logs.
7) Never throw any exception from finally block
try {
someMethod(); //Throws exceptionOne
} finally {
cleanUp(); //If finally also threw any exception the exceptionOne will be lost forever
}
This is fine, as long as cleanUp() can never throw any exception. In the above example, if someMethod() throws an exception, and in the finally block also, cleanUp() throws an exception, that second exception will come out of method and the original first exception (correct reason) will be lost forever. If the code that you call in a finally block can possibly throw an exception, make sure that you either handle it, or log it. Never let it come out of the finally block.
8) Always catch only those exceptions that you can actually handle
catch (NoSuchMethodException e) {
throw e; //Avoid this as it doesn't help anything
}
Well this is most important concept. Don’t catch any exception just for the sake of catching it. Catch any exception only if you want to handle it or, you want to provide additional contextual information in that exception. If you can’t handle it in catch block, then best advice is just don’t catch it only to re-throw it.
9) Don’t use printStackTrace() statement or similar methods
Never leave printStackTrace() after finishing your code. Chances are one of your fellow colleague will get one of those stack traces eventually, and have exactly zero knowledge as to what to do with it because it will not have any contextual information appended to it.
10) Use finally blocks instead of catch blocks if you are not going to handle exception
try {
someMethod(); //Method 2
} finally {
cleanUp(); //do cleanup here
}
This is also a good practice. If inside your method you are accessing some method 2, and method 2 throw some exception which you do not want to handle in method 1, but still want some cleanup in case exception occur, then do this cleanup in finally block. Do not use catch block.
11) Remember “Throw early catch late” principle
This is probably the most famous principle about Exception handling. It basically says that you should throw an exception as soon as you can, and catch it late as much as possible. You should wait until you have all the information to handle it properly.
This principle implicitly says that you will be more likely to throw it in the low-level methods, where you will be checking if single values are null or not appropriate. And you will be making the exception climb the stack trace for quite several levels until you reach a sufficient level of abstraction to be able to handle the problem.
12) Always clean up after handling the exception
If you are using resources like database connections or network connections, make sure you clean them up. If the API you are invoking uses only unchecked exceptions, you should still clean up resources after use, with try -- finally blocks. Inside try block access the resource and inside finally close the resource. Even if any exception occur in accessing the resource, then also resource will be closed gracefully.
13) Throw only relevant exception from a method
Relevancy is important to keep application clean. A method which tries to read a file; if throws NullPointerException then it will not give any relevant information to user. Instead it will be better if such exception is wrapped inside custom exception e.g. NoSuchFileFoundException then it will be more useful for users of that method.
14) Never use exceptions for flow control in your program
We have read it many times but sometimes we keep seeing code in our project where developer tries to use exceptions for application logic. Never do that. It makes code hard to read, understand and ugly.
15) Validate user input to catch adverse conditions very early in request processing
Always validate user input in very early stage, even before it reached to actual controller. It will help you to minimize the exception handling code in your core application logic. It also helps you in making application consistent if there is some error in user input.
For example: If in user registration application, you are following below logic:
1) Validate User
2) Insert User
3) Validate address
4) Insert address
5) If problem the Rollback everything
This is very incorrect approach. It can leave you database in inconsistent state in various scenarios. Rather validate everything in first place and then take the user data in dao layer and make DB updates. Correct approach is:
1) Validate User
2) Validate address
3) Insert User
4) Insert address
5) If problem the Rollback everything
16) Always include all information about an exception in single log message
LOGGER.debug(“Using cache sector A”);
LOGGER.debug(“Using retry sector B”);
Don’t do this.
Using a multi-line log message with multiple calls to LOGGER.debug() may look fine in your test case, but when it shows up in the log file of an app server with 400 threads running in parallel, all dumping information to the same log file, your two log messages may end up spaced out 1000 lines apart in the log file, even though they occur on subsequent lines in your code.
Do it like this:
LOGGER.debug(“Using cache sector A, using retry sector B”);
17) Pass all relevant information to exceptions to make them informative as much as possible
This is also very important to make exception messages and stack traces useful and informative. What is the use of a log, if you are not able to determine anything out of it. These type of logs just exist in your code for decoration purpose.
18) Always terminate the thread which it is interrupted
while (true) {
try {
Thread.sleep(100000);
} catch (InterruptedException e) {} //Don't do this
doSomethingCool();
}
InterruptedException is a clue to your code that it should stop whatever it’s doing. Some common use cases for a thread getting interrupted are the active transaction timing out, or a thread pool getting shut down. Instead of ignoring the InterruptedException, your code should do its best to finish up what it’s doing, and finish the current thread of execution. So to correct the example above:
while (true) {
try {
Thread.sleep(100000);
} catch (InterruptedException e) {
break;
}
}
doSomethingCool();
19) Use template methods for repeated try-catch
There is no use of having a similar catch block in 100 places in your code. It increases code duplicity which does not help anything. Use template methods for such cases.
For example below code tries to close a database connection.
class DBUtil{
public static void closeConnection(Connection conn){
try{
conn.close();
} catch(SQLException ex){
throw new RuntimeException("Cannot close connection", ex);
}
}
}
This type of method will be used in thousands of places in your application. Don’t put whole code in every place rather define above method and use it everywhere like below:
public void dataAccessCode() {
Connection conn = null;
try{
conn = getConnection();
....
} finally{
DBUtil.closeConnection(conn);
}
}
20) Document all exceptions in your application in javadoc
Make it a practice to javadoc all exceptions which a piece of code may throw at runtime. Also try to include possible course of action, user should follow in case these exception occur.
That’s all i have in my mind for now. If you found anything missing or you does not relate to my view on any point, drop me a comment. I will be happy to discuss.
