Classes

RLMArray is the container type in Realm used to define to-many relationships.

Unlike an NSArray, RLMArrays hold a single type, specified by the objectClassName property. This is referred to in these docs as the “type” of the array.

When declaring an RLMArray property, the type must be marked as conforming to a protocol by the same name as the objects it should contain (see the RLM_ARRAY_TYPE macro). In addition, the property can be declared using Objective‑C generics for better compile-time type safety.

RLM_ARRAY_TYPE(ObjectType)
...
@property RLMArray<ObjectType *><ObjectType> *arrayOfObjectTypes;

RLMArrays can be queried with the same predicates as RLMObject and RLMResults.

RLMArrays cannot be created directly. RLMArray properties on RLMObjects are lazily created when accessed, or can be obtained by querying a Realm.

Key-Value Observing

RLMArray supports array key-value observing on RLMArray properties on RLMObject subclasses, and the invalidated property on RLMArray instances themselves is key-value observing compliant when the RLMArray is attached to a managed RLMObject (RLMArrays on unmanaged RLMObjects will never become invalidated).

Because RLMArrays are attached to the object which they are a property of, they do not require using the mutable collection proxy objects from -mutableArrayValueForKey: or KVC-compatible mutation methods on the containing object. Instead, you can call the mutation methods on the RLMArray directly.

An RLMSortDescriptor stores a property name and a sort order for use with sortedResultsUsingDescriptors:. It is similar to NSSortDescriptor, but supports only the subset of functionality which can be efficiently run by Realm’s query engine.

RLMSortDescriptor instances are immutable.

A RLMCollectionChange object encapsulates information about changes to collections that are reported by Realm notifications.

RLMCollectionChange is passed to the notification blocks registered with -addNotificationBlock on RLMArray and RLMResults, and reports what rows in the collection changed since the last time the notification block was called.

The change information is available in two formats: a simple array of row indices in the collection for each type of change, and an array of index paths in a requested section suitable for passing directly to UITableView’s batch update methods. A complete example of updating a UITableView named tv:

[tv beginUpdates];
[tv deleteRowsAtIndexPaths:[changes deletionsInSection:0] withRowAnimation:UITableViewRowAnimationAutomatic];
[tv insertRowsAtIndexPaths:[changes insertionsInSection:0] withRowAnimation:UITableViewRowAnimationAutomatic];
[tv reloadRowsAtIndexPaths:[changes modificationsInSection:0] withRowAnimation:UITableViewRowAnimationAutomatic];
[tv endUpdates];

All of the arrays in an RLMCollectionChange are always sorted in ascending order.

RLMMigration instances encapsulate information intended to facilitate a schema migration.

A RLMMigration instance is passed into a user-defined RLMMigrationBlock block when updating the version of a Realm. This instance provides access to the old and new database schemas, the objects in the Realm, and provides functionality for modifying the Realm during the migration.

RLMObject is a base class for model objects representing data stored in Realms.

Define your model classes by subclassing RLMObject and adding properties to be persisted. Then instantiate and use your custom subclasses instead of using the RLMObject class directly.

// Dog.h
@interface Dog : RLMObject
@property NSString *name;
@property BOOL      adopted;
@end

// Dog.m
@implementation Dog
@end //none needed

Supported property types

• NSString
• NSInteger, int, long, float, and double
• BOOL or bool
• NSDate
• NSData
• NSNumber<X>, where X is one of RLMInt, RLMFloat, RLMDouble or RLMBool, for optional number properties
• RLMObject subclasses, to model many-to-one relationships.
• RLMArray<X>, where X is an RLMObject subclass, to model many-to-many relationships.

Querying

You can initiate queries directly via the class methods: allObjects, objectsWhere:, and objectsWithPredicate:. These methods allow you to easily query a custom subclass for instances of that class in the default Realm.

To search in a Realm other than the default Realm, use the allObjectsInRealm:, objectsInRealm:where:, and objectsInRealm:withPredicate: class methods.

Relationships

See our Cocoa guide for more details.

Key-Value Observing

All RLMObject properties (including properties you create in subclasses) are Key-Value Observing compliant, except for realm and objectSchema.

Keep the following tips in mind when observing Realm objects:

1. Unlike NSMutableArray properties, RLMArray properties do not require using the proxy object returned from -mutableArrayValueForKey:, or defining KVC mutation methods on the containing class. You can simply call methods on the RLMArray directly; any changes will be automatically observed by the containing object.
2. Unmanaged RLMObject instances cannot be added to a Realm while they have any observed properties.
3. Modifying managed RLMObjects within -observeValueForKeyPath:ofObject:change:context: is not recommended. Properties may change even when the Realm is not in a write transaction (for example, when -[RLMRealm refresh] is called after changes are made on a different thread), and notifications sent prior to the change being applied (when NSKeyValueObservingOptionPrior is used) may be sent at times when you cannot begin a write transaction.

This class represents Realm model object schemas.

When using Realm, RLMObjectSchema instances allow performing migrations and introspecting the database’s schema.

Object schemas map to tables in the core database.

RLMProperty instances represent properties persisted to Realm in the context of an object schema.

When using Realm, RLMProperty instances allow performing migrations and introspecting the database’s schema.

These property instances map to columns in the core database.

An RLMPropertyDescriptor instance represents a specific property on a given class.

An RLMRealm instance (also referred to as a Realm) represents a Realm database.

Realms can either be stored on disk (see +[RLMRealm realmWithURL:]) or in memory (see RLMRealmConfiguration).

RLMRealm instances are cached internally, and constructing equivalent RLMRealm objects (for example, by using the same path or identifier) multiple times on a single thread within a single iteration of the run loop will normally return the same RLMRealm object.

If you specifically want to ensure a RLMRealm instance is destroyed (for example, if you wish to open a Realm, check some property, and then possibly delete the Realm file and re-open it), place the code which uses the Realm within an @autoreleasepool {} and ensure you have no other strong references to it.

A token which is returned from methods which subscribe to changes to a Realm.

Change subscriptions in Realm return an RLMNotificationToken instance, which can be used to unsubscribe from the changes. You must store a strong reference to the token for as long as you want to continue to receive notifications. When you wish to stop, call the -stop method. Notifications are also stopped if the token is deallocated.

An RLMRealmConfiguration instance describes the different options used to create an instance of a Realm.

RLMRealmConfiguration instances are just plain NSObjects. Unlike RLMRealms and RLMObjects, they can be freely shared between threads as long as you do not mutate them.

Creating configuration objects for class subsets (by setting the objectClasses property) can be expensive. Because of this, you will normally want to cache and reuse a single configuration object for each distinct configuration rather than creating a new object each time you open a Realm.

RLMResults is an auto-updating container type in Realm returned from object queries. It represents the results of the query in the form of a collection of objects.

RLMResults can be queried using the same predicates as RLMObject and RLMArray, and you can chain queries to further filter results.

RLMResults always reflect the current state of the Realm on the current thread, including during write transactions on the current thread. The one exception to this is when using for...in fast enumeration, which will always enumerate over the objects which matched the query when the enumeration is begun, even if some of them are deleted or modified to be excluded by the filter during the enumeration.

RLMResults are lazily evaluated the first time they are accessed; they only run queries when the result of the query is requested. This means that chaining several temporary RLMResults to sort and filter your data does not perform any extra work processing the intermediate state.

Once the results have been evaluated or a notification block has been added, the results are eagerly kept up-to-date, with the work done to keep them up-to-date done on a background thread whenever possible.

RLMResults cannot be directly instantiated.

RLMLinkingObjects is an auto-updating container type. It represents a collection of objects that link to its parent object.

For more information, please see the Inverse Relationships section in the documentation.

RLMSchema instances represent collections of model object schemas persisted to a Realm.

When using Realm, RLMSchema instances allow performing migrations and introspecting the database’s schema.

Schemas map to collections of tables in the core database.