Use when setting up Firestore, designing schemas, doing CRUD, creating listeners, paginating queries, configuring indexes, enabling offline persistence, or writing security rules.
license
MIT
Firebase Cloud Firestore Skill
This skill defines how to correctly implement Cloud Firestore in Flutter applications, covering data modeling, queries, real-time updates, security rules, and scale optimization.
When to Use
Use this skill when:
Setting up and configuring Cloud Firestore in a Flutter project.
Designing document and collection structure or planning subcollections.
Performing read, write, batch, or transaction operations.
Implementing real-time listeners or paginated queries.
Optimizing for scale and avoiding write hotspots.
Writing or debugging Firestore security rules.
1. Database Selection
Choose Cloud Firestore when the app needs:
Rich, hierarchical data models with subcollections.
Complex queries: chaining filters, combining filtering and sorting on a property.
Transactions that atomically read and write data from any part of the database.
High availability (typical uptime 99.999%) or critical-level reliability.
Automatic scaling to millions of concurrent users.
Use Realtime Database instead for simple data models requiring simple lookups and extremely low-latency synchronization (typical response times under 10ms).
2. Setup and Configuration
flutter pub add cloud_firestore
import 'package:cloud_firestore/cloud_firestore.dart';
final db = FirebaseFirestore.instance; // after Firebase.initializeApp()
Location:
Select the database location closest to users and compute resources.
Use multi-region locations for critical apps (maximum availability and durability).
Use regional locations for lower costs and lower write latency.
iOS/macOS: Consider pre-compiled frameworks to improve build times:
pod 'FirebaseFirestore',
:git => 'https://github.com/invertase/firestore-ios-sdk-frameworks.git',
:tag => 'IOS_SDK_VERSION'
Offline persistence is enabled by default on mobile. Configure cache size:
Avoid document IDs . and .. (special meaning in Firestore paths).
Avoid forward slashes (/) in document IDs (path separators).
Do not use monotonically increasing document IDs (e.g., Customer1, Customer2) — causes write hotspots.
Use Firestore's automatic document IDs when possible:
final docRef = await db.collection("users").add({
'name': 'Ada Lovelace',
'email': 'ada@example.com',
'created_at': FieldValue.serverTimestamp(),
});
print('Created document with ID: ${docRef.id}');
Avoid these characters in field names (require extra escaping): .[]*`
Use subcollections within documents to organize complex, hierarchical data rather than deeply nested objects.
4. Indexing
Firestore queries are indexed by default; query performance is proportional to the result set size, not the dataset size.
Set collection-level index exemptions to reduce write latency and storage costs.
Disable Descending and Array indexing for fields that do not need them.
Exempt string fields with long values that are not used for querying.
Exempt fields with sequential values (e.g., timestamps) from indexing if not used in queries — avoids the 500 writes/second index limit.
Add single-field exemptions for TTL fields.
Exempt large array or map fields not used in queries — avoids the 40,000 index entries per document limit.
5. Read and Write Operations
Read All Documents in a Collection
final querySnapshot = await db.collection("users").get();
for (var doc in querySnapshot.docs) {
print("${doc.id} => ${doc.data()}");
}
Query with Filters
final query = db.collection("users")
.where("age", isGreaterThanOrEqualTo: 18)
.orderBy("age")
.limit(20);
final results = await query.get();
Cursor-Based Pagination
// First page
final first = db.collection("cities").orderBy("name").limit(25);
final firstSnapshot = await first.get();
// Next page using last document as cursor
final lastDoc = firstSnapshot.docs.last;
final next = db.collection("cities")
.orderBy("name")
.startAfterDocument(lastDoc)
.limit(25);
Do not use offsets for pagination — use cursors to avoid retrieving and being billed for skipped documents.