Docs Menu
Docs Home
/ /
Atlas Device SDKs
/ / /

CRUD - Read - Flutter SDK

On this page

  • Read Characteristics
  • Results Are Not Copies
  • Results Are Lazy
  • References Are Retained
  • Limiting Query Results
  • Pagination
  • Read Objects
  • Find Object by Primary Key
  • Query All Objects
  • Query Related Objects
  • Query Lists
  • Query Nested Collections of Mixed Data
  • Convert Lists or Sets to Results
  • Filter Results
  • Filter with Full-Text Search
  • Sort Results
  • Limit Results

You can read back the data that you have stored in the database by finding, filtering, and sorting objects.

A read from the database generally consists of the following steps:

  • Get all objects of a certain type from the database.

  • Optionally, filter the results.

Query operations return a results collection. These collections are live, meaning they always contain the latest results of the associated query.

Design your app's data access patterns around these three key read characteristics to read data as efficiently as possible.

Results to a query are not copies of your data. Modifying the results of a query modifies the data on disk directly. This memory mapping also means that results are live: that is, they always reflect the current state on disk.

The SDK only runs a query when you actually request the results of that query. This lazy evaluation enables you to write highly performant code for handling large data sets and complex queries. You can chain several filter operations without requiring extra work to process the intermediate state.

One benefit of the SDK's object model is that the SDK automatically retains all of an object's relationships as direct references. This enables you to traverse your graph of relationships directly through the results of a query.

A direct reference, or pointer, allows you to access a related object's properties directly through the reference.

Other databases typically copy objects from database storage into application memory when you need to work with them directly. Because application objects contain direct references, you are left with a choice: copy the object referred to by each direct reference out of the database in case it's needed, or just copy the foreign key for each object and query for the object with that key if it's accessed. If you choose to copy referenced objects into application memory, you can use up a lot of resources for objects that are never accessed, but if you choose to only copy the foreign key, referenced object lookups can cause your application to slow down.

The SDK bypasses all of this using zero-copy live objects. The SDK's object accessors point directly into database storage using memory mapping, so there is no distinction between the objects in the database and the results of your query in application memory. Because of this, you can traverse direct references across an entire database from any query result.

As a result of lazy evaluation, you do not need any special mechanism to limit query results with the SDK. For example, if your query matches thousands of objects, but you only want to load the first ten, simply access only the first ten elements of the results collection.

Thanks to lazy evaluation, the common task of pagination becomes quite simple. For example, suppose you have a results collection associated with a query that matches thousands of objects in your database. You display one hundred objects per page. To advance to any page, simply access the elements of the results collection starting at the index that corresponds to the target page.

Unless noted otherwise, the examples on this page use two object types, Person and Team.

class _Person {
late ObjectId id;
late String name;
late List<String> hobbies;
class _Team {
late ObjectId id;
late String name;
late List<_Person> crew;
late RealmValue eventLog;

Find an object by its primary key with Realm.find().

final luke = realm.find<Person>(lukePrimaryKey);

Retrieve a collection of all objects of a data model in the database with the Realm.all() method.

final people = realm.all<Person>();

New in version 1.8.0.

If your data model includes objects that reference other objects, you can query the relationship using the getBacklinks() method.

This method returns a results collection of all objects that link to the given object through a to-one, to-many, or inverse relationship. The following examples use the models defined on the Relationships page.

  • To-one relationship: In this example, we have a Bike object model with a to-one relationship with a Person object.

    We use the getBacklinks() method to find any Bike objects that link to the specified person through the owner property:

    // Persons have a to-one relationship with Bikes
    final person = realm.query<Person>("firstName == 'Anakin'").first;
    // Find all Bikes owned by a Person named 'Anakin'
    final allBikes = person.getBacklinks<Bike>('owner');
  • To-many relationship: In this example, we have a Scooter object model with a to-many relationship with a ScooterShop object.

    We use the getBacklinks() method to find any ScooterShops objects that link to the specified scooter through the scooters list property:

    // Scooters have a to-many relationship with ScooterShops
    final scooters = realm.query<Scooter>("name == 'Scooterbug'").first;
    // Find all ScooterShops with a Scooter named 'Scooterbug'
    final shops = scooters.getBacklinks<ScooterShop>('scooters');
  • Inverse relationship: In this example, we have a Task object model with an inverse relationship with a User object.

    We use the getBacklinks() method to find any User objects that link to the specified tasks through the tasks backlinks property:

    // Tasks have an inverse relationship to Users
    final inCompleteTasks = realm.query<Task>("isComplete == false");
    // Find all Users who have an incomplete Task
    for (final task in inCompleteTasks) {
    final ownersWithIncompleteTasks = task.getBacklinks<User>('tasks');
    for (final user in ownersWithIncompleteTasks) {
    print("User ${user.username} has incomplete tasks.");

You can query any list of RealmObjects or primitives.

final config = Configuration.local([Person.schema, Team.schema]);
final realm = Realm(config);
final heroes = Team(ObjectId(), 'Millenium Falcon Crew', crew: [
Person(ObjectId(), 'Luke'),
Person(ObjectId(), 'Leia'),
Person(ObjectId(), 'Han'),
Person(ObjectId(), 'Chewbacca')
realm.write(() => realm.add(heroes));
final lukeAndLeia = heroes.crew.query('name BEGINSWITH \$0', ['L']);

New in version 2.0.0.

In Flutter SDK v2.0.0 and later, RealmValue properties can contain collections (a list or map) of mixed data. These collections can be nested within collections and can contain other collections of mixed data.

You can query these using the same syntax as you would for a normal list or dictionary collection. Refer to the Realm Query Language documentation for more information on supported operators and list comparisons.

For nested collections, you can also use:

  • Bracket notation, which provides the following collection query operators:

    • [FIRST] and [LAST]: match the first or last elements within the collection.

    • [<int>]: match the element at the specific index.

    • [*]: match any element within the collection (this operator assumes a collection type at that path).

    • [SIZE]: match the collection length.

  • The @type operator, which supports the following values:

    • array and list: match a list collection.

    • dictionary and object: match a map collection.

    • collection: match a list or a map collection.

realm.write(() {
(Team(ObjectId(), 'Janitorial Staff',
eventLog: RealmValue.from({
'1': {
'date': DateTime.utc(5622, 8, 18, 12, 30, 0),
'type': ['work_order', 'maintenance'],
'summary': 'leaking pipes in control room',
'priority': 'high',
'2': {
'date': DateTime.utc(5622, 9, 18, 12, 30, 0),
'type': ['maintenance'],
'summary': 'trash compactor jammed',
'priority': 'low',
'comment': 'this is the second time this week'
(Team(ObjectId(), 'IT',
eventLog: RealmValue.from({
'1': {
'date': DateTime.utc(5622, 9, 20, 12, 30, 0),
'type': ['hardware', 'repair'],
'summary': 'lightsaber damage to server room',
'priority': 'high',
final teams = realm.all<Team>();
// Use bracket notation to query collection values at the specified path
final teamsWithHighPriorityEvents =
// Check any element at that path with [*]
teams.query("eventLog[*].priority == 'high'");
print(teamsWithHighPriorityEvents.length); // prints `2`
final teamsWithMaintenanceEvents =
// Check for the first element at that path with [FIRST]
teams.query("eventLog[*].type[FIRST] == 'maintenance'");
print(teamsWithMaintenanceEvents.length); // prints `1`
final teamsWithMultipleEvents =
// Check for collection at that path with matching elements
// Note that the order must match unless you use ANY or ALL
teams.query("eventLog[*].type[*] == {'maintenance', 'work_order'}");
teamsWithMultipleEvents.length); // prints `0` because order matters
final teamsWithEventsAsLists =
// Check the collection type with @type
teams.query("eventLog[*].type.@type == 'list'");
print(teamsWithEventsAsLists.length); // prints `2`

You can convert a RealmList or RealmSet to an instance of RealmResults using their respective methods:

These methods support lists and sets of RealmObjects as well as primitive values.

final config = Configuration.local([Person.schema, Team.schema]);
final realm = Realm(config);
final heroes = Team(ObjectId(), 'Millenium Falcon Crew', crew: [
Person(ObjectId(), 'Luke', hobbies: [
'Going to Tashi Station',
'Fixing the Moisture Vaporators'
Person(ObjectId(), 'Leia', hobbies: [
'Going on diplomatic missions',
'Rescuing short stormtroopers'
Person(ObjectId(), 'Han',
hobbies: ['Shooting first', 'Making fast Kessel Runs']),
Person(ObjectId(), 'Chewbacca', hobbies: [
'Fixing the Millenium Falcon',
'Tearing the arms off of droids'
realm.write(() => realm.add(heroes));
// Converts the Team object's 'crew' List into a RealmResults<Person>.
final heroesCrewAsResults = heroes.crew.asResults();
final luke = heroesCrewAsResults.query("name == 'Luke'").first;
// Converts Luke's 'hobbies' list into a RealmResults<String>
final lukeHobbiesAsResults = luke.hobbies.asResults();

Filter a RealmList to retrieve a specific segment of objects with the Realm.query() method. In the query() method's argument, use Realm Query Language to perform filtering. Realm Query Language is a string-based query language that you can use to retrieve objects from the database.

final team =
realm.query<Team>('name == \$0', ['Millennium Falcon Crew']).first;
final humanCrewMembers = team.crew.query('name != \$0', ['Chewbacca']);

You can use iterable arguments in your filter. For example:

final listOfNames = ['Luke', 'Leia'];
final matchingRealmObjects =
realm.query<Person>('name IN \$0', [listOfNames]);

You can also filter by inverse relationships using the @links.<Type>.<Property> syntax. For example, a filter can match a Task object based on properties of the User object that references it:

// Filter Tasks through the User's backlink property
// using `@links.<ObjectType>.<PropertyName>` syntax
final jarjarsIncompleteTasks = realm.query<Task>(
"ALL @links.User.tasks.username == 'jarjar_binks' AND isComplete == false");
final tasksForHan =
realm.query<Task>("ALL @links.User.tasks.username == 'han'");

For more information on constructing queries, refer to the Realm Query Language reference documentation.

You can use Realm Query Language (RQL) to query on properties that have a Full-Text Search Index (FTS) annotation. To query these properties, use the TEXT predicate in your query.

Exclude results for a word by placing the - character in front of the word. For example, a search for -sheep wool would include all search results for wool excluding those with sheep.

In Flutter SDK version 1.6.0 and later, you can specify prefixes by placing the * character at the end of a word. For example, wo* would include all search results for wool and woven. The Flutter SDK does not currently support suffix searches.

In the following example, we query on the Rug.pattern and Rug.material fields:

// Find rugs with a chevron pattern
final chevronRugs = realm.query<Rug>("pattern TEXT \$0", ["chevron"]);
// Find rugs with a wool material but not sheep wool
final notSheepWoolRugs = realm.query<Rug>("material TEXT \$0", [" -sheep wool"]);
// Find rugs with a material starting with "co-"
final coRugs = realm.query<Rug>("material TEXT \$0", ["co*"]);

Full-Text Search (FTS) indexes support:

  • Boolean match word searches, not searches for relevance.

  • Tokens are diacritics- and case-insensitive.

  • Tokens can only consist of characters from ASCII and the Latin-1 supplement (western languages).

  • All other characters are considered whitespace. Words split by a hyphen (-) like full-text are split into two tokens.

For more information on features of the FTS index, see the API reference for RealmIndexType.

Sort query results using the Realm Query Language SORT() operator in the query() method's argument.

Note that you can't use parameterized queries in RQL SORT() clauses. Instead, use strings or string interpolation.

realm.write(() {
Person(ObjectId(), 'Luke'),
Person(ObjectId(), 'Leia'),
Person(ObjectId(), 'Han'),
Person(ObjectId(), 'Chewbacca')
final alphabetizedPeople =
realm.query<Person>('TRUEPREDICATE SORT(name ASC)');
for (var person in alphabetizedPeople) {
// prints 'Chewbacca', 'Han', 'Leia', 'Luke'

Limit query results using the Realm Query Language LIMIT() operator in the query() method's argument.

Note that you can't use parameterized queries in RQL LIMIT() clauses. Instead, use strings or string interpolation.

realm.write(() {
Person(ObjectId(), 'Luke'),
Person(ObjectId(), 'Luke'),
Person(ObjectId(), 'Luke'),
Person(ObjectId(), 'Luke')
final limitedPeopleResults =
realm.query<Person>('name == \$0 LIMIT(2)', ['Luke']);
// prints `2`
← CRUD - Create - Flutter SDK