Creates, caches, and returns the default Firestore using the default FirebaseApp. Each subsequent invocation returns the same Firestore object.

Creates, caches, and returns the default Firestore object for the specified app. Each subsequent invocation returns the same Firestore object.

This method is in preview. API signature and functionality are subject to change.

Creates, caches, and returns named Firestore object for the specified FirebaseApp. Each subsequent invocation returns the same Firestore object.

This method is in preview. API signature and functionality are subject to change.

Creates, caches, and returns named Firestore object for the default app. Each subsequent invocation returns the same Firestore object.

Custom settings used to configure this Firestore object.

The Firebase App associated with this Firestore instance.

A PersistentCacheIndexManager which you can config persistent cache indexes used for local query execution.

NOTE: This preview method will be deprecated in a future major release. Consider using PersistentCacheIndexManager.enableIndexAutoCreation() to let the SDK decide whether to create cache indexes for queries running locally.

Configures indexing for local query execution. Any previous index configuration is overridden.

The index entries themselves are created asynchronously. You can continue to use queries that require indexing even if the indices are not yet available. Query execution will automatically start using the index once the index entries have been written.

The method accepts the JSON format exported by the Firebase CLI (firebase firestore:indexes). If the JSON format is invalid, the completion block will be invoked with an NSError.

NOTE: This preview method will be deprecated in a future major release. Consider using PersistentCacheIndexManager.enableIndexAutoCreation() to let the SDK decide whether to create cache indexes for queries running locally.

Configures indexing for local query execution. Any previous index configuration is overridden.

The index entries themselves are created asynchronously. You can continue to use queries that require indexing even if the indices are not yet available. Query execution will automatically start using the index once the index entries have been written.

Indexes are only supported with LevelDB persistence. Invoke set_persistence_enabled(true) before setting an index configuration. If LevelDB is not enabled, any index configuration will be rejected.

The method accepts the JSON format exported by the Firebase CLI (firebase firestore:indexes). If the JSON format is invalid, this method ignores the changes.

Gets a CollectionReference referring to the collection at the specified path within the database.

Gets a DocumentReference referring to the document at the specified path within the database.

Creates and returns a new Query that includes all documents in the database that are contained in a collection or subcollection with the given collectionID.

Executes the given updateBlock and then attempts to commit the changes applied within an atomic transaction.

The maximum number of writes allowed in a single transaction is 500, but note that each usage of FieldValue.serverTimestamp(), FieldValue.arrayUnion(), FieldValue.arrayRemove(), or FieldValue.increment() inside a transaction counts as an additional write.

In the updateBlock, a set of reads and writes can be performed atomically using the Transaction object passed to the block. After the updateBlock is run, Firestore will attempt to apply the changes to the server. If any of the data read has been modified outside of this transaction since being read, then the transaction will be retried by executing the updateBlock again. If the transaction still fails after 5 retries, then the transaction will fail.

Since the updateBlock may be executed multiple times, it should avoiding doing anything that would cause side effects.

Any value maybe be returned from the updateBlock. If the transaction is successfully committed, then the completion block will be passed that value. The updateBlock also has an NSErrorPointer out parameter. If this is set, then the transaction will not attempt to commit, and the given error will be passed to the completion block.

The Transaction object passed to the updateBlock contains methods for accessing documents and collections. Unlike other firestore access, data accessed with the transaction will not reflect local changes that have not been committed. For this reason, it is required that all reads are performed before any writes. Transactions must be performed while online. Otherwise, reads will fail, the final commit will fail, and the completion block will return an error.

Executes the given updateBlock and then attempts to commit the changes applied within an atomic transaction.

The maximum number of writes allowed in a single transaction is 500, but note that each usage of FieldValue.serverTimestamp(), FieldValue.arrayUnion(), FieldValue.arrayRemove(), or FieldValue.increment() inside a transaction counts as an additional write.

In the updateBlock, a set of reads and writes can be performed atomically using the Transaction object passed to the block. After the updateBlock is run, Firestore will attempt to apply the changes to the server. If any of the data read has been modified outside of this transaction since being read, then the transaction will be retried by executing the updateBlock again. If the transaction still fails after the attempting the number of times specified by the max_attempts property of the given TransactionOptions object, then the transaction will fail. If the given TransactionOptions is nil, then the default max_attempts of 5 will be used.

Since the updateBlock may be executed multiple times, it should avoiding doing anything that would cause side effects.

Any value maybe be returned from the updateBlock. If the transaction is successfully committed, then the completion block will be passed that value. The updateBlock also has an NSErrorPointer out parameter. If this is set, then the transaction will not attempt to commit, and the given error will be passed to the completion block.

The Transaction object passed to the updateBlock contains methods for accessing documents and collections. Unlike other firestore access, data accessed with the transaction will not reflect local changes that have not been committed. For this reason, it is required that all reads are performed before any writes. Transactions must be performed while online. Otherwise, reads will fail, the final commit will fail, and the completion block will return an error.

Creates a write batch, used for performing multiple writes as a single atomic operation.

The maximum number of writes allowed in a single batch is 500, but note that each usage of FieldValue.serverTimestamp(), FieldValue.arrayUnion(), FieldValue.arrayRemove(), or FieldValue.increment() inside a batch counts as an additional write.

Unlike transactions, write batches are persisted offline and therefore are preferable when you don’t need to condition your writes on read data.

Enables or disables logging from the Firestore client.

Configures Firestore to connect to an emulated host instead of the default remote backend. After Firestore has been used (i.e. a document reference has been instantiated), this value cannot be changed.

Re-enables usage of the network by this Firestore instance after a prior call to disableNetwork(completion:). Completion block, if provided, will be called once network uasge has been enabled.

Disables usage of the network by this Firestore instance. It can be re-enabled by via enableNetwork. While the network is disabled, any snapshot listeners or get calls will return results from cache and any write operations will be queued until the network is restored. The completion block, if provided, will be called once network usage has been disabled.

Clears the persistent storage. This includes pending writes and cached documents.

Must be called while the firestore instance is not started (after the app is shutdown or when the app is first initialized). On startup, this method must be called before other methods (other than Firestore.settings). If the firestore instance is still running, the function will complete with an error code of FailedPrecondition.

Note: clearPersistence(completion:) is primarily intended to help write reliable tests that use Firestore. It uses the most efficient mechanism possible for dropping existing data but does not attempt to securely overwrite or otherwise make cached data unrecoverable. For applications that are sensitive to the disclosure of cache data in between user sessions we strongly recommend not to enable persistence in the first place.

Waits until all currently pending writes for the active user have been acknowledged by the backend.

The completion block is called immediately without error if there are no outstanding writes. Otherwise, the completion block is called when all previously issued writes (including those written in a previous app session) have been acknowledged by the backend. The completion block does not wait for writes that were added after the method is called. If you wish to wait for additional writes, you have to call waitForPendingWrites again.

Any outstanding waitForPendingWrites(completion:) completion blocks are called with an error during user change.

Attaches a listener for a snapshots-in-sync event. The snapshots-in-sync event indicates that all listeners affected by a given change have fired, even if a single server-generated change affects multiple listeners.

NOTE: The snapshots-in-sync event only indicates that listeners are in sync with each other, but does not relate to whether those snapshots are in sync with the server. Use SnapshotMetadata in the individual listeners to determine if a snapshot is from the cache or the server.

Terminates this Firestore instance.

After calling terminate only the clearPersistence method may be used. Any other method will throw an error.

To restart after termination, simply create a new instance of Firestore with the firestore method.

Termination does not cancel any pending writes and any tasks that are awaiting a response from the server will not be resolved. The next time you start this instance, it will resume attempting to send these writes to the server.

Note: Under normal circumstances, calling this method is not required. This method is useful only when you want to force this instance to release all of its resources or in combination with clearPersistence to ensure that all local state is destroyed between test runs.

Loads a Firestore bundle into the local cache.

Loads a Firestore bundle into the local cache.

Loads a Firestore bundle into the local cache.

Loads a Firestore bundle into the local cache.

Reads a Query from the local cache, identified by the given name.

Named queries are packaged into bundles on the server side (along with the resulting documents) and loaded into local cache using loadBundle. Once in the local cache, you can use this method to extract a query by name.