In this post, I will first use an example to show what are instance variable initializer, instance initializer and static initializer, and then illustrate how the instance initializer works in Java.
1. Execution Order
Look at the following class, do you know which one gets executed first?
Output:
static initializer called instance initializer called constructor called instance initializer called constructor called
2. How does Java instance initializer work?
The instance initializer above contains a println statement. To understand how it works, we can treat it as a variable assignment statement, e.g., b = 0
. This can make it more obvious to understand.
Instead of
, you could write
Therefore, instance initializers and instance variable initializers are pretty much the same.
3. When are instance initializers useful?
The use of instance initializers are rare, but still it can be a useful alternative to instance variable initializers if:
(1) initializer code must handle exceptions
(2) perform calculations that can't be expressed with an instance variable initializer.
Of course, such code could be written in constructors. But if a class had multiple constructors, you would have to repeat the code in each constructor.
With an instance initializer, you can just write the code once, and it will be executed no matter what constructor is used to create the object. (I guess this is just a concept, and it is not used often.)
Another case in which instance initializers are useful is anonymous inner classes, which can't declare any constructors at all. (Will this be a good place to place a logging function?)
Thanks to Heinrich Hartmann for his comment:
Also note that Anonymous classes that implement interfaces [1] have no constructors. Therefore instance initializers are neede to execute any kinds of expressions at construction time.
Reference:
Object initialization in Java