How to create a Java application that uses Azure Cosmos DB for NoSQL and change feed processor
APPLIES TO: 
NoSQL
Azure Cosmos DB is a fully managed NoSQL database service provided by Azure. It allows you to build multiple-regionally distributed and highly scalable applications with ease. This how-to guide walks you through the process of creating a Java application that uses the Azure Cosmos DB for NoSQL database and implements the Change Feed Processor for real-time data processing. The Java application communicates with the Azure Cosmos DB for NoSQL using Azure Cosmos DB Java SDK v4.
Important
This tutorial is for Azure Cosmos DB Java SDK v4 only. Please view the Azure Cosmos DB Java SDK v4 Release notes, Maven repository, Change feed processor in Azure Cosmos DB, and Azure Cosmos DB Java SDK v4 troubleshooting guide for more information. If you are currently using an older version than v4, see the Migrate to Azure Cosmos DB Java SDK v4 guide for help upgrading to v4.
Prerequisites
Azure Cosmos DB Account: you can create it from the Azure portal or you can use Azure Cosmos DB Emulator as well.
Java Development Environment: Ensure you have Java Development Kit (JDK) installed on your machine with at least 8 version.
Azure Cosmos DB Java SDK V4: provides the necessary features to interact with Azure Cosmos DB.
Background
The Azure Cosmos DB change feed provides an event-driven interface to trigger actions in response to document insertion that has many uses.
The work of managing change feed events is largely taken care of by the change feed Processor library built into the SDK. This library is powerful enough to distribute change feed events among multiple workers, if that is desired. All you have to do is provide the change feed library a callback.
This simple example of Java application is demonstrating real-time data processing with Azure Cosmos DB and the Change Feed Processor. The application inserts sample documents into a "feed container" to simulate a data stream. The Change Feed Processor, bound to the feed container, processes incoming changes and logs the document content. The processor automatically manages leases for parallel processing.
Source code
You can clone the SDK example repo and find this example in SampleChangeFeedProcessor.java:
git clone https://github.com/Azure-Samples/azure-cosmos-java-sql-api-samples.git
cd azure-cosmos-java-sql-api-sample/src/main/java/com/azure/cosmos/examples/changefeed/
Walkthrough
- Configure the
ChangeFeedProcessorOptionsin a Java application using Azure Cosmos DB and Azure Cosmos DB Java SDK V4. TheChangeFeedProcessorOptionsprovides essential settings to control the behavior of the Change Feed Processor during data processing.
// Copyright (c) Microsoft Corporation. All rights reserved.
// Licensed under the MIT License.
package com.azure.cosmos.examples.changefeed;
import com.azure.cosmos.ChangeFeedProcessor;
import com.azure.cosmos.ChangeFeedProcessorBuilder;
import com.azure.cosmos.ConsistencyLevel;
import com.azure.cosmos.CosmosAsyncClient;
import com.azure.cosmos.CosmosAsyncContainer;
import com.azure.cosmos.CosmosAsyncDatabase;
import com.azure.cosmos.CosmosClientBuilder;
import com.azure.cosmos.CosmosException;
import com.azure.cosmos.examples.common.AccountSettings;
import com.azure.cosmos.examples.common.CustomPOJO2;
import com.azure.cosmos.implementation.Utils;
import com.azure.cosmos.models.ChangeFeedProcessorOptions;
import com.azure.cosmos.models.CosmosContainerProperties;
import com.azure.cosmos.models.CosmosContainerRequestOptions;
import com.azure.cosmos.models.CosmosContainerResponse;
import com.azure.cosmos.models.CosmosDatabaseResponse;
import com.azure.cosmos.models.ThroughputProperties;
import com.fasterxml.jackson.core.JsonProcessingException;
import com.fasterxml.jackson.databind.JsonNode;
import com.fasterxml.jackson.databind.ObjectMapper;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import reactor.core.scheduler.Schedulers;
import java.time.Duration;
import java.util.List;
import java.util.UUID;
import java.util.function.Consumer;
/**
* Sample for Change Feed Processor.
* This sample models an application where documents are being inserted into one container (the "feed container"),
* and meanwhile another worker thread or worker application is pulling inserted documents from the feed container's Change Feed
* and operating on them in some way. For one or more workers to process the Change Feed of a container, the workers must first contact the server
* and "lease" access to monitor one or more partitions of the feed container. The Change Feed Processor Library
* handles leasing automatically for you, however you must create a separate "lease container" where the Change Feed
* Processor Library can store and track leases container partitions.
*/
public class SampleChangeFeedProcessor {
public static int WAIT_FOR_WORK = 60000;
public static final String DATABASE_NAME = "db_" + UUID.randomUUID();
public static final String COLLECTION_NAME = "coll_" + UUID.randomUUID();
private static final ObjectMapper OBJECT_MAPPER = Utils.getSimpleObjectMapper();
protected static Logger logger = LoggerFactory.getLogger(SampleChangeFeedProcessor.class);
private static boolean isWorkCompleted = false;
private static ChangeFeedProcessorOptions options;
public static void main(String[] args) {
logger.info("Begin Sample");
try {
// <ChangeFeedProcessorOptions>
options = new ChangeFeedProcessorOptions();
options.setStartFromBeginning(false);
options.setLeasePrefix("myChangeFeedDeploymentUnit");
// </ChangeFeedProcessorOptions>
//Summary of the next four commands:
//-Create an asynchronous Azure Cosmos DB client and database so that we can issue async requests to the DB
//-Create a "feed container" and a "lease container" in the DB
logger.info("Create CosmosClient");
CosmosAsyncClient client = getCosmosClient();
logger.info("Create sample's database: " + DATABASE_NAME);
CosmosAsyncDatabase cosmosDatabase = createNewDatabase(client, DATABASE_NAME);
logger.info("Create container for documents: " + COLLECTION_NAME);
CosmosAsyncContainer feedContainer = createNewCollection(client, DATABASE_NAME, COLLECTION_NAME);
logger.info("Create container for lease: " + COLLECTION_NAME + "-leases");
CosmosAsyncContainer leaseContainer = createNewLeaseCollection(client, DATABASE_NAME, COLLECTION_NAME + "-leases");
//Model of a worker thread or application which leases access to monitor one or more feed container
//partitions via the Change Feed. In a real-world application you might deploy this code in an Azure function.
//The next line causes the worker to create and start an instance of the Change Feed Processor. See the implementation of getChangeFeedProcessor() for guidance
//on creating a handler for Change Feed events. In this stream, we also trigger the insertion of 10 documents on a separate
//thread.
// <StartChangeFeedProcessor>
logger.info("Start Change Feed Processor on worker (handles changes asynchronously)");
ChangeFeedProcessor changeFeedProcessorInstance = new ChangeFeedProcessorBuilder()
.hostName("SampleHost_1")
.feedContainer(feedContainer)
.leaseContainer(leaseContainer)
.handleChanges(handleChanges())
.buildChangeFeedProcessor();
changeFeedProcessorInstance.start()
.subscribeOn(Schedulers.boundedElastic())
.subscribe();
// </StartChangeFeedProcessor>
//These two lines model an application which is inserting ten documents into the feed container
logger.info("Start application that inserts documents into feed container");
createNewDocumentsCustomPOJO(feedContainer, 10, Duration.ofSeconds(3));
//This loop models the Worker main loop, which spins while its Change Feed Processor instance asynchronously
//handles incoming Change Feed events from the feed container. Of course in this sample, polling
//isWorkCompleted is unnecessary because items are being added to the feed container on the same thread, and you
//can see just above isWorkCompleted is set to true.
//But conceptually the worker is part of a different thread or application than the one which is inserting
//into the feed container; so this code illustrates the worker waiting and listening for changes to the feed container
long remainingWork = WAIT_FOR_WORK;
while (!isWorkCompleted && remainingWork > 0) {
Thread.sleep(100);
remainingWork -= 100;
}
//When all documents have been processed, clean up
if (isWorkCompleted) {
changeFeedProcessorInstance.stop().subscribe();
} else {
throw new RuntimeException("The change feed processor initialization and automatic create document feeding process did not complete in the expected time");
}
logger.info("Delete sample's database: " + DATABASE_NAME);
//deleteDatabase(cosmosDatabase);
Thread.sleep(500);
} catch (Exception e) {
e.printStackTrace();
}
logger.info("End Sample");
}
// <Delegate>
private static Consumer<List<JsonNode>> handleChanges() {
return (List<JsonNode> docs) -> {
logger.info("Start handleChanges()");
for (JsonNode document : docs) {
try {
//Change Feed hands the document to you in the form of a JsonNode
//As a developer you have two options for handling the JsonNode document provided to you by Change Feed
//One option is to operate on the document in the form of a JsonNode, as shown below. This is great
//especially if you do not have a single uniform data model for all documents.
logger.info("Document received: " + OBJECT_MAPPER.writerWithDefaultPrettyPrinter()
.writeValueAsString(document));
//You can also transform the JsonNode to a POJO having the same structure as the JsonNode,
//as shown below. Then you can operate on the POJO.
CustomPOJO2 pojo_doc = OBJECT_MAPPER.treeToValue(document, CustomPOJO2.class);
logger.info("id: " + pojo_doc.getId());
} catch (JsonProcessingException e) {
e.printStackTrace();
}
}
isWorkCompleted = true;
logger.info("End handleChanges()");
};
}
// </Delegate>
public static CosmosAsyncClient getCosmosClient() {
return new CosmosClientBuilder()
.endpoint(AccountSettings.HOST)
.key(AccountSettings.MASTER_KEY)
.contentResponseOnWriteEnabled(true)
.consistencyLevel(ConsistencyLevel.SESSION)
.buildAsyncClient();
}
public static CosmosAsyncDatabase createNewDatabase(CosmosAsyncClient client, String databaseName) {
CosmosDatabaseResponse databaseResponse = client.createDatabaseIfNotExists(databaseName).block();
return client.getDatabase(databaseResponse.getProperties().getId());
}
public static void deleteDatabase(CosmosAsyncDatabase cosmosDatabase) {
cosmosDatabase.delete().block();
}
public static CosmosAsyncContainer createNewCollection(CosmosAsyncClient client, String databaseName, String collectionName) {
CosmosAsyncDatabase databaseLink = client.getDatabase(databaseName);
CosmosAsyncContainer collectionLink = databaseLink.getContainer(collectionName);
CosmosContainerResponse containerResponse = null;
try {
containerResponse = collectionLink.read().block();
if (containerResponse != null) {
throw new IllegalArgumentException(String.format("Collection %s already exists in database %s.", collectionName, databaseName));
}
} catch (RuntimeException ex) {
if (ex instanceof CosmosException) {
CosmosException CosmosException = (CosmosException) ex;
if (CosmosException.getStatusCode() != 404) {
throw ex;
}
} else {
throw ex;
}
}
CosmosContainerProperties containerSettings = new CosmosContainerProperties(collectionName, "/pk");
CosmosContainerRequestOptions requestOptions = new CosmosContainerRequestOptions();
ThroughputProperties throughputProperties = ThroughputProperties.createManualThroughput(10000);
containerResponse = databaseLink.createContainer(containerSettings, throughputProperties, requestOptions).block();
if (containerResponse == null) {
throw new RuntimeException(String.format("Failed to create collection %s in database %s.", collectionName, databaseName));
}
return databaseLink.getContainer(containerResponse.getProperties().getId());
}
public static CosmosAsyncContainer createNewLeaseCollection(CosmosAsyncClient client, String databaseName, String leaseCollectionName) {
CosmosAsyncDatabase databaseLink = client.getDatabase(databaseName);
CosmosAsyncContainer leaseCollectionLink = databaseLink.getContainer(leaseCollectionName);
CosmosContainerResponse leaseContainerResponse = null;
try {
leaseContainerResponse = leaseCollectionLink.read().block();
if (leaseContainerResponse != null) {
leaseCollectionLink.delete().block();
try {
Thread.sleep(1000);
} catch (InterruptedException ex) {
ex.printStackTrace();
}
}
} catch (RuntimeException ex) {
if (ex instanceof CosmosException) {
CosmosException CosmosException = (CosmosException) ex;
if (CosmosException.getStatusCode() != 404) {
throw ex;
}
} else {
throw ex;
}
}
CosmosContainerProperties containerSettings = new CosmosContainerProperties(leaseCollectionName, "/id");
CosmosContainerRequestOptions requestOptions = new CosmosContainerRequestOptions();
ThroughputProperties throughputProperties = ThroughputProperties.createManualThroughput(400);
leaseContainerResponse = databaseLink.createContainer(containerSettings, throughputProperties, requestOptions).block();
if (leaseContainerResponse == null) {
throw new RuntimeException(String.format("Failed to create collection %s in database %s.", leaseCollectionName, databaseName));
}
return databaseLink.getContainer(leaseContainerResponse.getProperties().getId());
}
public static void createNewDocumentsCustomPOJO(CosmosAsyncContainer containerClient, int count, Duration delay) {
String suffix = UUID.randomUUID().toString();
for (int i = 0; i <= count; i++) {
CustomPOJO2 document = new CustomPOJO2();
document.setId(String.format("0%d-%s", i, suffix));
document.setPk(document.getId()); // This is a very simple example, so we'll just have a partition key (/pk) field that we set equal to id
containerClient.createItem(document).subscribe(doc -> {
logger.info("Document write: " + doc);
});
long remainingWork = delay.toMillis();
try {
while (remainingWork > 0) {
Thread.sleep(100);
remainingWork -= 100;
}
} catch (InterruptedException iex) {
// exception caught
break;
}
}
}
}
- Initialize
ChangeFeedProcessorwith relevant configurations, including the host name, feed container, lease container, and data handling logic. Thestart()method initiates the data processing, enabling concurrent and real-time processing of incoming data changes from the feed container.
// Copyright (c) Microsoft Corporation. All rights reserved.
// Licensed under the MIT License.
package com.azure.cosmos.examples.changefeed;
import com.azure.cosmos.ChangeFeedProcessor;
import com.azure.cosmos.ChangeFeedProcessorBuilder;
import com.azure.cosmos.ConsistencyLevel;
import com.azure.cosmos.CosmosAsyncClient;
import com.azure.cosmos.CosmosAsyncContainer;
import com.azure.cosmos.CosmosAsyncDatabase;
import com.azure.cosmos.CosmosClientBuilder;
import com.azure.cosmos.CosmosException;
import com.azure.cosmos.examples.common.AccountSettings;
import com.azure.cosmos.examples.common.CustomPOJO2;
import com.azure.cosmos.implementation.Utils;
import com.azure.cosmos.models.ChangeFeedProcessorOptions;
import com.azure.cosmos.models.CosmosContainerProperties;
import com.azure.cosmos.models.CosmosContainerRequestOptions;
import com.azure.cosmos.models.CosmosContainerResponse;
import com.azure.cosmos.models.CosmosDatabaseResponse;
import com.azure.cosmos.models.ThroughputProperties;
import com.fasterxml.jackson.core.JsonProcessingException;
import com.fasterxml.jackson.databind.JsonNode;
import com.fasterxml.jackson.databind.ObjectMapper;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import reactor.core.scheduler.Schedulers;
import java.time.Duration;
import java.util.List;
import java.util.UUID;
import java.util.function.Consumer;
/**
* Sample for Change Feed Processor.
* This sample models an application where documents are being inserted into one container (the "feed container"),
* and meanwhile another worker thread or worker application is pulling inserted documents from the feed container's Change Feed
* and operating on them in some way. For one or more workers to process the Change Feed of a container, the workers must first contact the server
* and "lease" access to monitor one or more partitions of the feed container. The Change Feed Processor Library
* handles leasing automatically for you, however you must create a separate "lease container" where the Change Feed
* Processor Library can store and track leases container partitions.
*/
public class SampleChangeFeedProcessor {
public static int WAIT_FOR_WORK = 60000;
public static final String DATABASE_NAME = "db_" + UUID.randomUUID();
public static final String COLLECTION_NAME = "coll_" + UUID.randomUUID();
private static final ObjectMapper OBJECT_MAPPER = Utils.getSimpleObjectMapper();
protected static Logger logger = LoggerFactory.getLogger(SampleChangeFeedProcessor.class);
private static boolean isWorkCompleted = false;
private static ChangeFeedProcessorOptions options;
public static void main(String[] args) {
logger.info("Begin Sample");
try {
// <ChangeFeedProcessorOptions>
options = new ChangeFeedProcessorOptions();
options.setStartFromBeginning(false);
options.setLeasePrefix("myChangeFeedDeploymentUnit");
// </ChangeFeedProcessorOptions>
//Summary of the next four commands:
//-Create an asynchronous Azure Cosmos DB client and database so that we can issue async requests to the DB
//-Create a "feed container" and a "lease container" in the DB
logger.info("Create CosmosClient");
CosmosAsyncClient client = getCosmosClient();
logger.info("Create sample's database: " + DATABASE_NAME);
CosmosAsyncDatabase cosmosDatabase = createNewDatabase(client, DATABASE_NAME);
logger.info("Create container for documents: " + COLLECTION_NAME);
CosmosAsyncContainer feedContainer = createNewCollection(client, DATABASE_NAME, COLLECTION_NAME);
logger.info("Create container for lease: " + COLLECTION_NAME + "-leases");
CosmosAsyncContainer leaseContainer = createNewLeaseCollection(client, DATABASE_NAME, COLLECTION_NAME + "-leases");
//Model of a worker thread or application which leases access to monitor one or more feed container
//partitions via the Change Feed. In a real-world application you might deploy this code in an Azure function.
//The next line causes the worker to create and start an instance of the Change Feed Processor. See the implementation of getChangeFeedProcessor() for guidance
//on creating a handler for Change Feed events. In this stream, we also trigger the insertion of 10 documents on a separate
//thread.
// <StartChangeFeedProcessor>
logger.info("Start Change Feed Processor on worker (handles changes asynchronously)");
ChangeFeedProcessor changeFeedProcessorInstance = new ChangeFeedProcessorBuilder()
.hostName("SampleHost_1")
.feedContainer(feedContainer)
.leaseContainer(leaseContainer)
.handleChanges(handleChanges())
.buildChangeFeedProcessor();
changeFeedProcessorInstance.start()
.subscribeOn(Schedulers.boundedElastic())
.subscribe();
// </StartChangeFeedProcessor>
//These two lines model an application which is inserting ten documents into the feed container
logger.info("Start application that inserts documents into feed container");
createNewDocumentsCustomPOJO(feedContainer, 10, Duration.ofSeconds(3));
//This loop models the Worker main loop, which spins while its Change Feed Processor instance asynchronously
//handles incoming Change Feed events from the feed container. Of course in this sample, polling
//isWorkCompleted is unnecessary because items are being added to the feed container on the same thread, and you
//can see just above isWorkCompleted is set to true.
//But conceptually the worker is part of a different thread or application than the one which is inserting
//into the feed container; so this code illustrates the worker waiting and listening for changes to the feed container
long remainingWork = WAIT_FOR_WORK;
while (!isWorkCompleted && remainingWork > 0) {
Thread.sleep(100);
remainingWork -= 100;
}
//When all documents have been processed, clean up
if (isWorkCompleted) {
changeFeedProcessorInstance.stop().subscribe();
} else {
throw new RuntimeException("The change feed processor initialization and automatic create document feeding process did not complete in the expected time");
}
logger.info("Delete sample's database: " + DATABASE_NAME);
//deleteDatabase(cosmosDatabase);
Thread.sleep(500);
} catch (Exception e) {
e.printStackTrace();
}
logger.info("End Sample");
}
// <Delegate>
private static Consumer<List<JsonNode>> handleChanges() {
return (List<JsonNode> docs) -> {
logger.info("Start handleChanges()");
for (JsonNode document : docs) {
try {
//Change Feed hands the document to you in the form of a JsonNode
//As a developer you have two options for handling the JsonNode document provided to you by Change Feed
//One option is to operate on the document in the form of a JsonNode, as shown below. This is great
//especially if you do not have a single uniform data model for all documents.
logger.info("Document received: " + OBJECT_MAPPER.writerWithDefaultPrettyPrinter()
.writeValueAsString(document));
//You can also transform the JsonNode to a POJO having the same structure as the JsonNode,
//as shown below. Then you can operate on the POJO.
CustomPOJO2 pojo_doc = OBJECT_MAPPER.treeToValue(document, CustomPOJO2.class);
logger.info("id: " + pojo_doc.getId());
} catch (JsonProcessingException e) {
e.printStackTrace();
}
}
isWorkCompleted = true;
logger.info("End handleChanges()");
};
}
// </Delegate>
public static CosmosAsyncClient getCosmosClient() {
return new CosmosClientBuilder()
.endpoint(AccountSettings.HOST)
.key(AccountSettings.MASTER_KEY)
.contentResponseOnWriteEnabled(true)
.consistencyLevel(ConsistencyLevel.SESSION)
.buildAsyncClient();
}
public static CosmosAsyncDatabase createNewDatabase(CosmosAsyncClient client, String databaseName) {
CosmosDatabaseResponse databaseResponse = client.createDatabaseIfNotExists(databaseName).block();
return client.getDatabase(databaseResponse.getProperties().getId());
}
public static void deleteDatabase(CosmosAsyncDatabase cosmosDatabase) {
cosmosDatabase.delete().block();
}
public static CosmosAsyncContainer createNewCollection(CosmosAsyncClient client, String databaseName, String collectionName) {
CosmosAsyncDatabase databaseLink = client.getDatabase(databaseName);
CosmosAsyncContainer collectionLink = databaseLink.getContainer(collectionName);
CosmosContainerResponse containerResponse = null;
try {
containerResponse = collectionLink.read().block();
if (containerResponse != null) {
throw new IllegalArgumentException(String.format("Collection %s already exists in database %s.", collectionName, databaseName));
}
} catch (RuntimeException ex) {
if (ex instanceof CosmosException) {
CosmosException CosmosException = (CosmosException) ex;
if (CosmosException.getStatusCode() != 404) {
throw ex;
}
} else {
throw ex;
}
}
CosmosContainerProperties containerSettings = new CosmosContainerProperties(collectionName, "/pk");
CosmosContainerRequestOptions requestOptions = new CosmosContainerRequestOptions();
ThroughputProperties throughputProperties = ThroughputProperties.createManualThroughput(10000);
containerResponse = databaseLink.createContainer(containerSettings, throughputProperties, requestOptions).block();
if (containerResponse == null) {
throw new RuntimeException(String.format("Failed to create collection %s in database %s.", collectionName, databaseName));
}
return databaseLink.getContainer(containerResponse.getProperties().getId());
}
public static CosmosAsyncContainer createNewLeaseCollection(CosmosAsyncClient client, String databaseName, String leaseCollectionName) {
CosmosAsyncDatabase databaseLink = client.getDatabase(databaseName);
CosmosAsyncContainer leaseCollectionLink = databaseLink.getContainer(leaseCollectionName);
CosmosContainerResponse leaseContainerResponse = null;
try {
leaseContainerResponse = leaseCollectionLink.read().block();
if (leaseContainerResponse != null) {
leaseCollectionLink.delete().block();
try {
Thread.sleep(1000);
} catch (InterruptedException ex) {
ex.printStackTrace();
}
}
} catch (RuntimeException ex) {
if (ex instanceof CosmosException) {
CosmosException CosmosException = (CosmosException) ex;
if (CosmosException.getStatusCode() != 404) {
throw ex;
}
} else {
throw ex;
}
}
CosmosContainerProperties containerSettings = new CosmosContainerProperties(leaseCollectionName, "/id");
CosmosContainerRequestOptions requestOptions = new CosmosContainerRequestOptions();
ThroughputProperties throughputProperties = ThroughputProperties.createManualThroughput(400);
leaseContainerResponse = databaseLink.createContainer(containerSettings, throughputProperties, requestOptions).block();
if (leaseContainerResponse == null) {
throw new RuntimeException(String.format("Failed to create collection %s in database %s.", leaseCollectionName, databaseName));
}
return databaseLink.getContainer(leaseContainerResponse.getProperties().getId());
}
public static void createNewDocumentsCustomPOJO(CosmosAsyncContainer containerClient, int count, Duration delay) {
String suffix = UUID.randomUUID().toString();
for (int i = 0; i <= count; i++) {
CustomPOJO2 document = new CustomPOJO2();
document.setId(String.format("0%d-%s", i, suffix));
document.setPk(document.getId()); // This is a very simple example, so we'll just have a partition key (/pk) field that we set equal to id
containerClient.createItem(document).subscribe(doc -> {
logger.info("Document write: " + doc);
});
long remainingWork = delay.toMillis();
try {
while (remainingWork > 0) {
Thread.sleep(100);
remainingWork -= 100;
}
} catch (InterruptedException iex) {
// exception caught
break;
}
}
}
}
- Specify the delegate handles incoming data changes using the
handleChanges()method. The method processes the received JsonNode documents from the Change Feed. As a developer you have two options for handling the JsonNode document provided to you by Change Feed. One option is to operate on the document in the form of a JsonNode. This is great especially if you don't have a single uniform data model for all documents. The second option - transform the JsonNode to a POJO having the same structure as the JsonNode. Then you can operate on the POJO.
// Copyright (c) Microsoft Corporation. All rights reserved.
// Licensed under the MIT License.
package com.azure.cosmos.examples.changefeed;
import com.azure.cosmos.ChangeFeedProcessor;
import com.azure.cosmos.ChangeFeedProcessorBuilder;
import com.azure.cosmos.ConsistencyLevel;
import com.azure.cosmos.CosmosAsyncClient;
import com.azure.cosmos.CosmosAsyncContainer;
import com.azure.cosmos.CosmosAsyncDatabase;
import com.azure.cosmos.CosmosClientBuilder;
import com.azure.cosmos.CosmosException;
import com.azure.cosmos.examples.common.AccountSettings;
import com.azure.cosmos.examples.common.CustomPOJO2;
import com.azure.cosmos.implementation.Utils;
import com.azure.cosmos.models.ChangeFeedProcessorOptions;
import com.azure.cosmos.models.CosmosContainerProperties;
import com.azure.cosmos.models.CosmosContainerRequestOptions;
import com.azure.cosmos.models.CosmosContainerResponse;
import com.azure.cosmos.models.CosmosDatabaseResponse;
import com.azure.cosmos.models.ThroughputProperties;
import com.fasterxml.jackson.core.JsonProcessingException;
import com.fasterxml.jackson.databind.JsonNode;
import com.fasterxml.jackson.databind.ObjectMapper;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import reactor.core.scheduler.Schedulers;
import java.time.Duration;
import java.util.List;
import java.util.UUID;
import java.util.function.Consumer;
/**
* Sample for Change Feed Processor.
* This sample models an application where documents are being inserted into one container (the "feed container"),
* and meanwhile another worker thread or worker application is pulling inserted documents from the feed container's Change Feed
* and operating on them in some way. For one or more workers to process the Change Feed of a container, the workers must first contact the server
* and "lease" access to monitor one or more partitions of the feed container. The Change Feed Processor Library
* handles leasing automatically for you, however you must create a separate "lease container" where the Change Feed
* Processor Library can store and track leases container partitions.
*/
public class SampleChangeFeedProcessor {
public static int WAIT_FOR_WORK = 60000;
public static final String DATABASE_NAME = "db_" + UUID.randomUUID();
public static final String COLLECTION_NAME = "coll_" + UUID.randomUUID();
private static final ObjectMapper OBJECT_MAPPER = Utils.getSimpleObjectMapper();
protected static Logger logger = LoggerFactory.getLogger(SampleChangeFeedProcessor.class);
private static boolean isWorkCompleted = false;
private static ChangeFeedProcessorOptions options;
public static void main(String[] args) {
logger.info("Begin Sample");
try {
// <ChangeFeedProcessorOptions>
options = new ChangeFeedProcessorOptions();
options.setStartFromBeginning(false);
options.setLeasePrefix("myChangeFeedDeploymentUnit");
// </ChangeFeedProcessorOptions>
//Summary of the next four commands:
//-Create an asynchronous Azure Cosmos DB client and database so that we can issue async requests to the DB
//-Create a "feed container" and a "lease container" in the DB
logger.info("Create CosmosClient");
CosmosAsyncClient client = getCosmosClient();
logger.info("Create sample's database: " + DATABASE_NAME);
CosmosAsyncDatabase cosmosDatabase = createNewDatabase(client, DATABASE_NAME);
logger.info("Create container for documents: " + COLLECTION_NAME);
CosmosAsyncContainer feedContainer = createNewCollection(client, DATABASE_NAME, COLLECTION_NAME);
logger.info("Create container for lease: " + COLLECTION_NAME + "-leases");
CosmosAsyncContainer leaseContainer = createNewLeaseCollection(client, DATABASE_NAME, COLLECTION_NAME + "-leases");
//Model of a worker thread or application which leases access to monitor one or more feed container
//partitions via the Change Feed. In a real-world application you might deploy this code in an Azure function.
//The next line causes the worker to create and start an instance of the Change Feed Processor. See the implementation of getChangeFeedProcessor() for guidance
//on creating a handler for Change Feed events. In this stream, we also trigger the insertion of 10 documents on a separate
//thread.
// <StartChangeFeedProcessor>
logger.info("Start Change Feed Processor on worker (handles changes asynchronously)");
ChangeFeedProcessor changeFeedProcessorInstance = new ChangeFeedProcessorBuilder()
.hostName("SampleHost_1")
.feedContainer(feedContainer)
.leaseContainer(leaseContainer)
.handleChanges(handleChanges())
.buildChangeFeedProcessor();
changeFeedProcessorInstance.start()
.subscribeOn(Schedulers.boundedElastic())
.subscribe();
// </StartChangeFeedProcessor>
//These two lines model an application which is inserting ten documents into the feed container
logger.info("Start application that inserts documents into feed container");
createNewDocumentsCustomPOJO(feedContainer, 10, Duration.ofSeconds(3));
//This loop models the Worker main loop, which spins while its Change Feed Processor instance asynchronously
//handles incoming Change Feed events from the feed container. Of course in this sample, polling
//isWorkCompleted is unnecessary because items are being added to the feed container on the same thread, and you
//can see just above isWorkCompleted is set to true.
//But conceptually the worker is part of a different thread or application than the one which is inserting
//into the feed container; so this code illustrates the worker waiting and listening for changes to the feed container
long remainingWork = WAIT_FOR_WORK;
while (!isWorkCompleted && remainingWork > 0) {
Thread.sleep(100);
remainingWork -= 100;
}
//When all documents have been processed, clean up
if (isWorkCompleted) {
changeFeedProcessorInstance.stop().subscribe();
} else {
throw new RuntimeException("The change feed processor initialization and automatic create document feeding process did not complete in the expected time");
}
logger.info("Delete sample's database: " + DATABASE_NAME);
//deleteDatabase(cosmosDatabase);
Thread.sleep(500);
} catch (Exception e) {
e.printStackTrace();
}
logger.info("End Sample");
}
// <Delegate>
private static Consumer<List<JsonNode>> handleChanges() {
return (List<JsonNode> docs) -> {
logger.info("Start handleChanges()");
for (JsonNode document : docs) {
try {
//Change Feed hands the document to you in the form of a JsonNode
//As a developer you have two options for handling the JsonNode document provided to you by Change Feed
//One option is to operate on the document in the form of a JsonNode, as shown below. This is great
//especially if you do not have a single uniform data model for all documents.
logger.info("Document received: " + OBJECT_MAPPER.writerWithDefaultPrettyPrinter()
.writeValueAsString(document));
//You can also transform the JsonNode to a POJO having the same structure as the JsonNode,
//as shown below. Then you can operate on the POJO.
CustomPOJO2 pojo_doc = OBJECT_MAPPER.treeToValue(document, CustomPOJO2.class);
logger.info("id: " + pojo_doc.getId());
} catch (JsonProcessingException e) {
e.printStackTrace();
}
}
isWorkCompleted = true;
logger.info("End handleChanges()");
};
}
// </Delegate>
public static CosmosAsyncClient getCosmosClient() {
return new CosmosClientBuilder()
.endpoint(AccountSettings.HOST)
.key(AccountSettings.MASTER_KEY)
.contentResponseOnWriteEnabled(true)
.consistencyLevel(ConsistencyLevel.SESSION)
.buildAsyncClient();
}
public static CosmosAsyncDatabase createNewDatabase(CosmosAsyncClient client, String databaseName) {
CosmosDatabaseResponse databaseResponse = client.createDatabaseIfNotExists(databaseName).block();
return client.getDatabase(databaseResponse.getProperties().getId());
}
public static void deleteDatabase(CosmosAsyncDatabase cosmosDatabase) {
cosmosDatabase.delete().block();
}
public static CosmosAsyncContainer createNewCollection(CosmosAsyncClient client, String databaseName, String collectionName) {
CosmosAsyncDatabase databaseLink = client.getDatabase(databaseName);
CosmosAsyncContainer collectionLink = databaseLink.getContainer(collectionName);
CosmosContainerResponse containerResponse = null;
try {
containerResponse = collectionLink.read().block();
if (containerResponse != null) {
throw new IllegalArgumentException(String.format("Collection %s already exists in database %s.", collectionName, databaseName));
}
} catch (RuntimeException ex) {
if (ex instanceof CosmosException) {
CosmosException CosmosException = (CosmosException) ex;
if (CosmosException.getStatusCode() != 404) {
throw ex;
}
} else {
throw ex;
}
}
CosmosContainerProperties containerSettings = new CosmosContainerProperties(collectionName, "/pk");
CosmosContainerRequestOptions requestOptions = new CosmosContainerRequestOptions();
ThroughputProperties throughputProperties = ThroughputProperties.createManualThroughput(10000);
containerResponse = databaseLink.createContainer(containerSettings, throughputProperties, requestOptions).block();
if (containerResponse == null) {
throw new RuntimeException(String.format("Failed to create collection %s in database %s.", collectionName, databaseName));
}
return databaseLink.getContainer(containerResponse.getProperties().getId());
}
public static CosmosAsyncContainer createNewLeaseCollection(CosmosAsyncClient client, String databaseName, String leaseCollectionName) {
CosmosAsyncDatabase databaseLink = client.getDatabase(databaseName);
CosmosAsyncContainer leaseCollectionLink = databaseLink.getContainer(leaseCollectionName);
CosmosContainerResponse leaseContainerResponse = null;
try {
leaseContainerResponse = leaseCollectionLink.read().block();
if (leaseContainerResponse != null) {
leaseCollectionLink.delete().block();
try {
Thread.sleep(1000);
} catch (InterruptedException ex) {
ex.printStackTrace();
}
}
} catch (RuntimeException ex) {
if (ex instanceof CosmosException) {
CosmosException CosmosException = (CosmosException) ex;
if (CosmosException.getStatusCode() != 404) {
throw ex;
}
} else {
throw ex;
}
}
CosmosContainerProperties containerSettings = new CosmosContainerProperties(leaseCollectionName, "/id");
CosmosContainerRequestOptions requestOptions = new CosmosContainerRequestOptions();
ThroughputProperties throughputProperties = ThroughputProperties.createManualThroughput(400);
leaseContainerResponse = databaseLink.createContainer(containerSettings, throughputProperties, requestOptions).block();
if (leaseContainerResponse == null) {
throw new RuntimeException(String.format("Failed to create collection %s in database %s.", leaseCollectionName, databaseName));
}
return databaseLink.getContainer(leaseContainerResponse.getProperties().getId());
}
public static void createNewDocumentsCustomPOJO(CosmosAsyncContainer containerClient, int count, Duration delay) {
String suffix = UUID.randomUUID().toString();
for (int i = 0; i <= count; i++) {
CustomPOJO2 document = new CustomPOJO2();
document.setId(String.format("0%d-%s", i, suffix));
document.setPk(document.getId()); // This is a very simple example, so we'll just have a partition key (/pk) field that we set equal to id
containerClient.createItem(document).subscribe(doc -> {
logger.info("Document write: " + doc);
});
long remainingWork = delay.toMillis();
try {
while (remainingWork > 0) {
Thread.sleep(100);
remainingWork -= 100;
}
} catch (InterruptedException iex) {
// exception caught
break;
}
}
}
}
- Build and run the Java application. The application starts the Change Feed Processor, insert sample documents into the feed container, and process the incoming changes.
Conclusion
In this guide, you learned how to create a Java application using Azure Cosmos DB Java SDK V4 that uses the Azure Cosmos DB for NoSQL database and uses the Change Feed Processor for real-time data processing. You can extend this application to handle more complex use cases and build robust, scalable, and multiple-regionally distributed applications using Azure Cosmos DB.
Additional resources
Next steps
You can now proceed to learn more about change feed estimator in the following articles: