Sprinkles
Sprinkles is a boiler-plate-reduction-library for dealing with databases in android applications. Some would call it a kind of ORM but I don't see it that way. Sprinkles lets SQL do what it is good at, making complex queries. SQL however is a mess (in my opinion) when is comes to everything else. This is why sprinkles helps you with things such as inserting, updating, and destroying models. Sprinkles will also help you with the tedious task of unpacking a cursor into a model. Sprinkles actively supports version 2.3 of Android and above but it should work on older versions as well.
Download
Using gradle, add the following to your build.gradle
. Just replace x.x.x
with the correct version of the library (found under the releases tab).
dependencies {
compile 'se.emilsjolander:sprinkles:x.x.x'
}
If you are not using gradle for whatever reason i suggest you clone the repository and check out the latest tag.
Getting started
When you have added the library to your project add a model class to it. I will demonstrate this with a Note.java
class. I have omitted the import statements to keep it brief.
@Table("Notes")
public class Note extends Model {
@Key
@AutoIncrement
@Column("id")
private long id;
@Column("title")
public String title;
@Column("body")
public String body;
public long getId() {
return id;
}
}
Ok, a lot of important stuff in this short class. First of all, a model must subclass se.emilsjolander.sprinkles.Model
and it also must have a @Table
annotations specifying the table name that the model corresponds to. After the class declaration we have declared three members: id
, title
and body
. Notice how all of them have a @Column
annotation to mark that they are not only a member of this class but also a column of the table that this class represents. We have one last annotation in the above example. The @AutoIncrement
annotation tells sprinkles that the field should automatically be set upon the creation of its corresponding row in the table. Key columns are the columns that are used to decide whether a model is already stored in the database when using methods such as delete()
and save()
.
Before using this class you must migrate it into the database. I recommend doing this in the onCreate()
method of an Application
subclass like this:
public class MyApplication extends Application {
@Override
public void onCreate() {
super.onCreate();
Sprinkles sprinkles = Sprinkles.init(getApplicationContext());
sprinkles.addMigration(new Migration() {
@Override
protected void onPreMigrate() {
// do nothing
}
@Override
protected void doMigration(SQLiteDatabase db) {
db.execSQL(
"CREATE TABLE Notes (" +
"id INTEGER PRIMARY KEY AUTOINCREMENT,"+
"title TEXT,"+
"body TEXT"+
")"
);
}
@Override
protected void onPostMigrate() {
// do nothing
}
});
}
}
Now you can happily create new instances of this class and save it to the database like so:
public void saveStuff() {
Note n = new Note();
n.title = "Sprinkles is awesome!";
n.body = "yup, sure is!";
n.save(); // when this call finishes n.getId() will return a valid id
}
You can also query for this note like this:
public void queryStuff() {
Note n = Query.one(Note.class, "select * from Notes where title=?", "Sprinkles is awesome!").get();
}
There is a lot more you can do with sprinkles so please read the next section which covers the whole API!
API
###Annotations
@Table
Used to associate a model class with a SQL table.@AutoIncrement
Used to mark a field as an auto-incrementing. The field must be anint
or along
.@Column
Used to associate a class field with a SQL column.@DynamicColumn
Used to associate a class field with a dynamic SQL column such as an alias in a query.@Key
Used to mark a field as a key. Multiple keys in a class are allowed and will result in a composite key. Keys will most often want to be mapped directly to primary keys in your database.
###Saving The save method is both an insert and an update method, the correct operation will be done depending on the model's existence in the database. The first two methods below are synchronous, the second is for use together with a transaction (more on that later). There are also two asynchronous methods, one with a callback and one without. The synchronous methods will return a boolean indicating if the model was saved or not. The asynchronous method with a callback will just not invoke the callback if saving failed.
boolean save();
boolean save(Transaction t);
void saveAsync();
void saveAsync(OnSavedCallback callback);
All the save methods use this method to check if a model exists in the database. You are free to use it as well.
boolean exists();
###Deleting Similar to saving there are four methods that let you delete a model. These work in the same way as save but will not return a boolean indicating the result.
void delete();
void delete(Transaction t);
void deleteAsync();
void deleteAsync(OnDeletedCallback callback);
###Querying Start a query with on of the following static methods:
Query.one(Class<? extends QueryResult> clazz, String sql, Object[] args);
Query.many(Class<? extends QueryResult> clazz, String sql, Object[] args);
Query.all(Class<? extends Model> clazz);
Notice that unlike androids built in query methods you can send in an array of objects instead of an array of strings.
Once the query has been started you can get the result with two different methods:
<T extends QueryResult> get();
boolean getAsync(LoaderManager lm, ResultHandler<? extends Model> handler, Class<? extends Model>... respondsToUpdatedOf);
get()
returns either the QueryResult
or a list of the QueryResult
represented by the Class
you sent in as the first argument to the query method. getAsync()
is the same only that the result is delivered on a callback function after executing get()
on another thread. getAsync()
also delivers updated results once the backing model of the query is updated if you return true
indicating you want further updates. getAsync()
uses loaders and therefore needs a LoaderManager
instance. getAsync()
also takes an optional array of classes which is used when the query relies on more models than the one you are querying for and you want the query to be updated when those models change as well.
###CursorList
All Queries return a CursorList
subclass. This is a Iterable
subclass which lazily unpacks a cursor into its corresponding model when you ask for the next item. This leads to having the efficiency of a Cursor
but without the pain. Excluding the Iterable
methods CursorList
also provides the following methods.
public int size();
public T get(int pos);
public List<T> asList();
Remember to always call close()
on a CursorList
instance! This will close the underlying cursor.
###ModelList
For mass saving/deletion of models you can use the ModelList
class. It extends ArrayList
and has the following additional methods:
public static <E extends Model> ModelList<E> from(CursorList<E> cursorList);
public boolean saveAll();
public boolean saveAll(Transaction t);
public void saveAllAsync();
public void saveAllAsync(OnAllSavedCallback callback);
public void deleteAll();
public void deleteAll(Transaction t);
public void deleteAllAsync();
public void deleteAllAsync(OnAllDeletedCallback callback);
from(CursorList<E extends Model> cursorList)
is a helper method which creates a ModelList
from a CursorList
, so you can e.g. delete all models from a previous query in one batch. Be aware, that the cursor is not closed for you when calling this method and you have to do it yourself!
###Transactions
Both save()
and delete()
methods exist which take in a Transaction
. Here is a quick example on how to use them. If any exception is thrown while saving a model or if any model fails to save the transaction will be rolled back.
public void doTransaction(List<Note> notes) {
Transaction t = new Transaction();
try {
for (Note n : notes) {
if (!n.save(t)) {
return;
}
}
t.setSuccessful(true);
} finally {
t.finish();
}
}
###Callbacks Each model subclass can override a couple of callbacks.
Use the following callback to ensure that your model is not saved in an invalid state.
@Override
public boolean isValid() {
// check model validity
}
Use the following callback to update a variable before the model is created
@Override
protected void beforeCreate() {
mCreatedAt = System.currentTimeMillis();
}
Use the following callback to update a variable before the model is saved. This is called directly before beforeCreate()
if the model is saved for the first time.
@Override
protected void beforeSave() {
mUpdatedAt = System.currentTimeMillis();
}
Use the following callback to clean up things related to the model but not stored in the database. Perhaps a file on the internal storage?
@Override
protected void afterDelete() {
// clean up some things?
}
###Migrations
Migrations are the way you add things to your database. I suggest putting all your migrations in the onCreate()
method of a Application
subclass. Here is a quick example of how that would look:
public class MyApplication extends Application {
@Override
public void onCreate() {
super.onCreate();
Sprinkles sprinkles = Sprinkles.init(getApplicationContext());
sprinkles.addMigration(new Migration() {
@Override
protected void onPreMigrate() {
// do nothing
}
@Override
protected void doMigration(SQLiteDatabase db) {
db.execSQL(
"CREATE TABLE Notes (" +
"id INTEGER PRIMARY KEY AUTOINCREMENT,"+
"title TEXT,"+
"body TEXT"+
")"
);
}
@Override
protected void onPostMigrate() {
// do nothing
}
});
}
}
Migrations are performed using raw SQL, this allowes full freedom to use all of the powerfull contraints that are possible to put on columns. Two optional methods are provided that allow you do some form of processing of your data before and after a migration, this can be usefull when recreating a table with different properties but you want to keep the data that was previously stored in the now deleted table. Once a migration has been added with sprinkles.addMigration()
it should NEVER be changed, and all new migrations should be added after the previous migration. This ensures both old and new clients will have a consistent database and you will not need to care about database versioning.
###Type serializers
Through an instance of Sprinkles
you can register your own TypeSerializer
instances via registerType()
for serializing an object in your model into a column in the database. Sprinkles uses a TypeSerializer
implementation internally for all the different data types that it supports. So check out the se.emilsjolander.sprinkles.typeserializers
package for example implementations. These serializers will be used both when saving a model and when querying rows from the database.
###ContentObservers Sprinkles supports ContentObservers for change notifications. By registering your models for observation you can ensure your ContentObserver will be notified of changes.
SprinklesContentObserver observer;
ContentObserver myCustomObserver = ...;
this.observer = new SprinklesContentObserver(myCustomObserver);
@Override
public void onResume() {
super.onResume();
this.observer.register(Note.class, true); // true/false for notify descendants
}
@Override
public void onPause() {
super.onPause();
this.observer.unregister();
}
###Relationships Sprinkles does nothing to handle relationships for you; this is by design. You will have to use the regular ways to handle relationships in SQL. Sprinkles gives you all the tools needed for this and it works very well.