What is the main purpose of setTag() getTag() methods of View?

I'd like to add few words.

Although using get/setTag(Object) seems to be very useful in the particular case of a ViewHolder pattern, I'd recommend to think twice before using it in other cases. There is almost always another solution with better design.

The main reason is that code like that becomes unsupportable pretty quickly.

• It is non-obvious for other developers what you designed to store as tag in the view. The methods setTag/getTag are not descriptive at all.

• It just stores an Object, which requires to be cast when you want to getTag. You can get unexpected crashes later when you decide to change the type of stored object in the tag.

• Here's a real-life story: We had a pretty big project with a lot of adapters, async operations with views and so on. One developer decided to set/getTag in his part of code, but another one had already set the tag to this view. In the end, someone couldn't find his own tag and was very confused. It cost us several hours to find the bug.

setTag(int key, Object tag) looks much better, cause you can generate unique keys for every tag (using id resources), but there is a significant restriction for Android < 4.0. From Lint docs:

Prior to Android 4.0, the implementation of View.setTag(int, Object) would store the objects in a static map, where the values were strongly referenced. This means that if the object contains any references pointing back to the context, the context (which points to pretty much everything else) will leak. If you pass a view, the view provides a reference to the context that created it. Similarly, view holders typically contain a view, and cursors are sometimes also associated with views.

We can use setTag() and getTag() to set and get custom objects as per our requirement. The setTag() method takes an argument of type Object, and getTag() returns an Object.

For example,

Person p = new Person();
p.setName("Ramkailash");
p.setId(2000001);
button1.setTag(p);

Let's say you generate a bunch of views that are similar. You could set an OnClickListener for each view individually:

button1.setOnClickListener(new OnClickListener ... );
button2.setOnClickListener(new OnClickListener ... );
 ...

Then you have to create a unique onClick method for each view even if they do the similar things, like:

public void onClick(View v) {
    doAction(1); // 1 for button1, 2 for button2, etc.
}

This is because onClick has only one parameter, a View, and it has to get other information from instance variables or final local variables in enclosing scopes. What we really want is to get information from the views themselves.

Enter getTag/setTag:

button1.setTag(1);
button2.setTag(2);

Now we can use the same OnClickListener for every button:

listener = new OnClickListener() {
    @Override
    public void onClick(View v) {
        doAction(v.getTag());
    }
};

It's basically a way for views to have memories.