I\'m working on an android application, that uses a navigation drawer to switch between two fragments. However, each time I switch, the fragment is completely recreated.
this is a little late response. if you're using view pager for fragments, set the off screen page limit of the fragment to the number of fragments created.
mViewPager.setOffscreenPageLimit(3); // number of fragments here is 3
Use the attach/detach method with tags:
Detach will destroy the view hirachy but keeps the state, like if on the backstack; this will let the "not-visible" fragment have a smaller memory footprint. But mind you that you need to correctly implement the fragment lifecycle (which you should do in the first place)
Detach the given fragment from the UI. This is the same state as when it is put on the back stack: the fragment is removed from the UI, however its state is still being actively managed by the fragment manager. When going into this state its view hierarchy is destroyed.
The first time you add the fragment
FragmentTransaction t = getSupportFragmentManager().beginTransaction();
t.add(android.R.id.content, new MyFragment(),MyFragment.class.getSimpleName());
t.commit();
then you detach it
FragmentTransaction t = getSupportFragmentManager().beginTransaction();
t.detach(MyFragment.class.getSimpleName());
t.commit();
and attach it again if switched back, state will be kept
FragmentTransaction t = getSupportFragmentManager().beginTransaction();
t.attach(getSupportFragmentManager().findFragmentByTag(MyFragment.class.getSimpleName()));
t.commit();
But you always have to check if the fragment was added yet, if not then add it, else just attach it:
if (getSupportFragmentManager().findFragmentByTag(MyFragment.class.getSimpleName()) == null) {
FragmentTransaction t = getSupportFragmentManager().beginTransaction();
t.add(android.R.id.content, new MyFragment(), MyFragment.class.getSimpleName());
t.commit();
} else {
FragmentTransaction t = getSupportFragmentManager().beginTransaction();
t.attach(getSupportFragmentManager().findFragmentByTag(MyFragment.class.getSimpleName()));
t.commit();
}
I did this before like this:
if (mPrevFrag != fragment) {
// Change
FragmentTransaction ft = fragmentManager.beginTransaction();
if (mPrevFrag != null){
ft.hide(mPrevFrag);
}
ft.show(fragment);
ft.commit();
mPrevFrag = fragment;
}
(you will need to track your pervious fragment in this solution)
I guess you can not directly manipulate the lifecycle mechanisms of your Fragments
. The very fact that you can findFragmentByTag
is not very bad. It means that the Fragment object is not recreated fully, if it is already commited. The existing Fragment
just passes all the lifecycle steps each Fragment
has - that means that only UI is "recreated".
It is a very convenient and useful memory management strategy - and appropriate, in most cases. Fragment
which is gone, has the resources which have to be utilized in order to de-allocate memory.
If you just cease using this strategy, the memory usage of your application could increase badly.
Nonetheless, there are retained fragments, which lifecycle is a bit different and do not correspond to the Activity they are attached to. Typically, they are used to retain some things you want to save, for example, to manage configuration changes
However, the fragment [re]creation strategy depends on the context - that is, what you would like to solve, and what are the trade-offs that you are willing to accept.