Paging library overview   Part of Android Jetpack.

The Paging Library helps you load and display small chunks of data at a time. Loading partial data on demand reduces usage of network bandwidth and system resources.

This guide provides several conceptual examples of the library, along with an overview of how it works. To view complete examples of how this library functions, try out the codelab and samples from the additional resources section.

Library architecture

This section describes and shows the main components of the paging library.

PagedList

The Paging Library's key component is the PagedList class, which loads chunks of your app's data, or pages. As more data is needed, it's paged into the existing PagedList object. If any loaded data changes, a new instance of PagedList is emitted to the observable data holder from a LiveData or RxJava2-based object. As PagedList objects are generated, your app's UI presents their contents, all while respecting your UI controllers' lifecycles.

The following code snippet shows how you can configure your app's view model to load and present data using a LiveData holder of PagedList objects:

Kotlin

class ConcertViewModel(concertDao: ConcertDao) : ViewModel() {
    val concertList: LiveData<PagedList<Concert>> =
            concertDao.concertsByDate().toLiveData(pageSize = 50)
}

Java

public class ConcertViewModel extends ViewModel {
    private ConcertDao concertDao;
    public final LiveData<PagedList<Concert>> concertList;

    // Creates a PagedList object with 50 items per page.
    public ConcertViewModel(ConcertDao concertDao) {
        this.concertDao = concertDao;
        concertList = new LivePagedListBuilder<>(
                concertDao.concertsByDate(), 50).build();
    }
}

Data

Each instance of PagedList loads an up-to-date snapshot of your app's data from its corresponding DataSource object. Data flows from your app's backend or database into the PagedList object.

The following example uses the Room persistence library to organize your app's data, but if you want to store your data using another means, you can also provide your own data source factory.

Kotlin

@Dao
interface ConcertDao {
    // The Int type parameter tells Room to use a PositionalDataSource object.
    @Query("SELECT * FROM concerts ORDER BY date DESC")
    fun concertsByDate(): DataSource.Factory<Int, Concert>
}

Java

@Dao
public interface ConcertDao {
    // The Integer type parameter tells Room to use a
    // PositionalDataSource object.
    @Query("SELECT * FROM concerts ORDER BY date DESC")
    DataSource.Factory<Integer, Concert> concertsByDate();
}

To learn more about how you can load data into PagedList objects, see the guide on how to Load paged data.

UI

The PagedList class works with a PagedListAdapter to load items into a RecyclerView. These classes work together to fetch and display content as it's loaded, prefetching out-of-view content and animating content changes.

To learn more, see the guide on how to Display paged lists.

Support different data architectures

The Paging Library supports the following data architectures:

• Served only from a backend server.
• Stored only in an on-device database.
• A combination of the other sources, using the on-device database as a cache.

Figure 1 shows how data flows in each of these architecture scenarios. In the case of a network-only or database-only solution, the data flows directly to your app's UI model. If you're using a combined approach, data flows from your backend server, into an on-device database, and then to your app's UI model. Every once in a while, the endpoint of each data flow runs out of data to load, at which point it requests more data from the component that provided the data. For example, when an on-device database runs out of data, it requests more data from the server.

Figure 1. How data flows through each of the architectures that the Paging Library supports

The remainder of this section provides recommendations for configuring each data flow use case.

Network only

To display data from a backend server, use the synchronous version of the Retrofit API to load information into your own custom DataSource object.

Note: The Paging Library's DataSource objects don't provide any error handling because different apps handle and present error UIs in different ways. If an error occurs, defer to your result callback, and retry the request later. See the PagingWithNetwork sample for an example of this behavior.

Database only

Set up your RecyclerView to observe local storage, preferably using the Room persistence library. That way, whenever data is inserted or modified in your app's database, these changes are automatically reflected in the RecyclerView that's displaying this data.

Network and database

After you've started observing the database, you can listen for when the database is out of data by using PagedList.BoundaryCallback. You can then fetch more items from your network and insert them into the database. If your UI is observing the database, that's all you need to do.

Handle network errors

When using a network to fetch or page the data that you're displaying using the Paging Library, it's important to not treat the network as being either "available" or "unavailable" all the time, as many connections are intermittent or flaky:

• A particular server might fail to respond to a network request.
• The device might be connected to a network that's slow or weak.

Instead, your app should check each request for failures and recover as gracefully as possible in cases where the network isn't available. For example, you can provide a "retry" button for users to select if the data refresh step doesn't work. If an error occurs during the data paging step, it's best to retry the paging requests automatically.

Update your existing app

If your app already consumes data from a database or a backend source, it's possible to upgrade directly to functionality that the Paging Library provides. This section shows how to upgrade an app that has a common existing design.

Custom paging solutions

If you use custom functionality to load small subsets of data from your app's data source, you can replace this logic with that from the PagedList class. Instances of PagedList offer built-in connections to common data sources. These instances also provide adapters for RecyclerView objects that you might include in your app's UI.

Data loaded using lists instead of pages

If you use an in-memory list as the backing data structure for your UI's adapter, consider observing data updates using a PagedList class if the number of items in the list can become large. Instances of PagedList can use either LiveData<PagedList> or Observable<List> to pass data updates to your app's UI, minimizing load times and memory usage. Better still, replacing a List object with a PagedList object in your app doesn't require any changes to your app's UI structure or data updating logic.

Associate a data cursor with a list view using CursorAdapter

Your app might use a CursorAdapter to associate data from a Cursor with a ListView. In that case, you usually need to migrate from a ListView to a RecyclerView as your app's list UI container, then replace the Cursor component with either Room or PositionalDataSource, depending on whether instances of Cursor access a SQLite database.

In some situations, such as when working with instances of Spinner, you provide only the adapter itself. A library then takes the data that's loaded into that adapter and displays the data for you. In these situations, change the type of your adapter's data to LiveData<PagedList>, then wrap this list in an ArrayAdapter object before attempting to have a library class inflate these items in a UI.

Load content asynchronously using AsyncListUtil

If you're using AsyncListUtil objects to load and display groups of information asynchronously, the Paging Library lets you load data more easily:

• Your data doesn't need to be positional. The Paging Library lets you load data directly from your backend using keys that the network provides.
• Your data can be uncountably large. Using the Paging Library, you can load data into pages until there isn't any data remaining.
• You can observe your data more easily. The Paging library can present your data that your app's ViewModel holds in an observable data structure.

Note: If your existing app accesses a SQLite database, see the section on using the Room persistence library.

Database examples

The following code snippets show several possible ways of having all the pieces work together.

Observing paged data using LiveData

The following code snippet shows all the pieces working together. As concert events are added, removed, or changed in the database, the content in the RecyclerView is automatically and efficiently updated:

Kotlin

@Dao
interface ConcertDao {
    // The Int type parameter tells Room to use a PositionalDataSource
    // object, with position-based loading under the hood.
    @Query("SELECT * FROM concerts ORDER BY date DESC")
    fun concertsByDate(): DataSource.Factory<Int, Concert>
}

class ConcertViewModel(concertDao: ConcertDao) : ViewModel() {
    val concertList: LiveData<PagedList<Concert>> =
            concertDao.concertsByDate().toLiveData(pageSize = 50)
}

class ConcertActivity : AppCompatActivity() {
    public override fun onCreate(savedInstanceState: Bundle?) {
        super.onCreate(savedInstanceState)
        val viewModel = ViewModelProviders.of(this)
                .get<ConcertViewModel>()
        val recyclerView = findViewById(R.id.concert_list)
        val adapter = ConcertAdapter()
        viewModel.livePagedList.observe(this, PagedList(adapter::submitList))
        recyclerView.setAdapter(adapter)
    }
}

class ConcertAdapter() :
        PagedListAdapter<Concert, ConcertViewHolder>(DIFF_CALLBACK) {
    fun onBindViewHolder(holder: ConcertViewHolder, position: Int) {
        val concert: Concert? = getItem(position)

        // Note that "concert" is a placeholder if it's null.
        holder.bindTo(concert)
    }

    companion object {
        private val DIFF_CALLBACK = object :
                DiffUtil.ItemCallback<Concert>() {
            // Concert details may have changed if reloaded from the database,
            // but ID is fixed.
            override fun areItemsTheSame(oldConcert: Concert,
                    newConcert: Concert) = oldConcert.id == newConcert.id

            override fun areContentsTheSame(oldConcert: Concert,
                    newConcert: Concert) = oldConcert == newConcert
        }
    }
}

Java

@Dao
public interface ConcertDao {
    // The Integer type parameter tells Room to use a PositionalDataSource
    // object, with position-based loading under the hood.
    @Query("SELECT * FROM concerts ORDER BY date DESC")
    DataSource.Factory<Integer, Concert> concertsByDate();
}

public class ConcertViewModel extends ViewModel {
    private ConcertDao concertDao;
    public final LiveData<PagedList<Concert>> concertList;

    public ConcertViewModel(ConcertDao concertDao) {
        this.concertDao = concertDao;
        concertList = new LivePagedListBuilder<>(
            concertDao.concertsByDate(), /* page size */ 50).build();
    }
}

public class ConcertActivity extends AppCompatActivity {
    @Override
    public void onCreate(@Nullable Bundle savedInstanceState) {
        super.onCreate(savedInstanceState);
        ConcertViewModel viewModel =
                ViewModelProviders.of(this).get(ConcertViewModel.class);
        RecyclerView recyclerView = findViewById(R.id.concert_list);
        ConcertAdapter adapter = new ConcertAdapter();
        viewModel.concertList.observe(this, adapter::submitList);
        recyclerView.setAdapter(adapter);
    }
}

public class ConcertAdapter
        extends PagedListAdapter<Concert, ConcertViewHolder> {
    protected ConcertAdapter() {
        super(DIFF_CALLBACK);
    }

    @Override
    public void onBindViewHolder(@NonNull ConcertViewHolder holder,
            int position) {
        Concert concert = getItem(position);
        if (concert != null) {
            holder.bindTo(concert);
        } else {
            // Null defines a placeholder item - PagedListAdapter automatically
            // invalidates this row when the actual object is loaded from the
            // database.
            holder.clear();
        }
    }

    private static DiffUtil.ItemCallback<Concert> DIFF_CALLBACK =
            new DiffUtil.ItemCallback<Concert>() {
        // Concert details may have changed if reloaded from the database,
        // but ID is fixed.
        @Override
        public boolean areItemsTheSame(Concert oldConcert, Concert newConcert) {
            return oldConcert.getId() == newConcert.getId();
        }

        @Override
        public boolean areContentsTheSame(Concert oldConcert,
                Concert newConcert) {
            return oldConcert.equals(newConcert);
        }
    };
}

Observing paged data using RxJava2

If you prefer using RxJava2 instead of LiveData, you can instead create an Observable or Flowable object:

Kotlin

class ConcertViewModel(concertDao: ConcertDao) : ViewModel() {
    val concertList: Observable<PagedList<Concert>> =
            concertDao.concertsByDate().toObservable(pageSize = 50)
}

Java

public class ConcertViewModel extends ViewModel {
    private ConcertDao concertDao;
    public final Observable<PagedList<Concert>> concertList;

    public ConcertViewModel(ConcertDao concertDao) {
        this.concertDao = concertDao;

        concertList = new RxPagedListBuilder<>(
                concertDao.concertsByDate(), /* page size */ 50)
                        .buildObservable();
    }
}

You can then start and stop observing the data using the code in the following snippet:

Kotlin

class ConcertActivity : AppCompatActivity() {
    private val adapter: ConcertAdapter()
    private lateinit var viewModel: ConcertViewModel

    private val disposable = CompositeDisposable()

    public override fun onCreate(savedInstanceState: Bundle?) {
        super.onCreate(savedInstanceState)
        val recyclerView = findViewById(R.id.concert_list)
        viewModel = ViewModelProviders.of(this)
                .get<ConcertViewModel>()
        recyclerView.setAdapter(adapter)
    }

    override fun onStart() {
        super.onStart()
        disposable.add(viewModel.concertList
                .subscribe(adapter::submitList)))
    }

    override fun onStop() {
        super.onStop()
        disposable.clear()
    }
}

Java

public class ConcertActivity extends AppCompatActivity {
    private ConcertAdapter adapter = new ConcertAdapter();
    private ConcertViewModel viewModel;

    private CompositeDisposable disposable = new CompositeDisposable();

    @Override
    public void onCreate(@Nullable Bundle savedInstanceState) {
        super.onCreate(savedInstanceState);
        RecyclerView recyclerView = findViewById(R.id.concert_list);

        viewModel = ViewModelProviders.of(this).get(ConcertViewModel.class);
        recyclerView.setAdapter(adapter);
    }

    @Override
    protected void onStart() {
        super.onStart();
        disposable.add(viewModel.concertList
                .subscribe(adapter.submitList(flowableList)
        ));
    }

    @Override
    protected void onStop() {
        super.onStop();
        disposable.clear();
    }
}

The code for the ConcertDao and ConcertAdapter are the same for an RxJava2-based solution as they are for a LiveData-based solution.

Provide feedback

Share your feedback and ideas with us through these resources:

Issue tracker

Report issues so we can fix bugs.

Additional resources

To learn more about the Paging Library, consult the following resources.

Samples

• Android Architecture Components Paging sample
• Paging With Network Sample

Codelabs

• Android Paging codelab

Videos

• Android Jetpack: manage infinite lists with RecyclerView and Paging (Google I/O '18)
• Android Jetpack: Paging