# Work with JSON documents with the Enhanced Document API for DynamoDB
The [Enhanced Document API](https://sdk.amazonaws.com/java/api/latest/software/amazon/awssdk/enhanced/dynamodb/document/package-summary.html) for AWS SDK for Java 2.x is designed to work with document-oriented data that has no fixed schema. However, it also lets you use custom classes to map individual attributes.
The Enhanced Document API is the successor to the [Document API](https://docs.aws.amazon.com/AWSJavaSDK/latest/javadoc/com/amazonaws/services/dynamodbv2/document/DynamoDB.html) of the AWS SDK for Java v1.x.
**Contents**
+ [Get started using the Enhanced Document API](ddb-en-client-doc-api-steps.md)
+ [Create a `DocumentTableSchema` and a `DynamoDbTable`](ddb-en-client-doc-api-steps.md#ddb-en-client-doc-api-steps-createschema)
+ [Build enhanced documents](ddb-en-client-doc-api-steps-create-ed.md)
+ [Build from a JSON string](ddb-en-client-doc-api-steps-create-ed.md#ddb-en-client-doc-api-steps-create-ed-fromJson)
+ [Build from individual elements](ddb-en-client-doc-api-steps-create-ed.md#ddb-en-client-doc-api-steps-create-ed-fromparts)
+ [Perform CRUD operations](ddb-en-client-doc-api-steps-use.md)
+ [Access enhanced document attributes as custom objects](ddb-en-client-doc-api-convert.md)
+ [Use an `EnhancedDocument` without DynamoDB](ddb-en-client-doc-api-standalone.md)
# Get started using the Enhanced Document API
The Enhanced Document API requires the same [dependencies](ddb-en-client-getting-started.md#ddb-en-client-gs-dep) that are needed for the DynamoDB Enhanced Client API. It also requires a [`DynamoDbEnhancedClient` instance](ddb-en-client-getting-started-dynamodbTable.md#ddb-en-client-getting-started-dynamodbTable-eclient) as shown at the start of this topic.
Because the Enhanced Document API was released with version 2.20.3 of the AWS SDK for Java 2.x, you need that version or greater.
## Create a `DocumentTableSchema` and a `DynamoDbTable`
To invoke commands against a DynamoDB table using the Enhanced Document API, associate the table with a client-side [DynamoDbTable](https://sdk.amazonaws.com/java/api/latest/software/amazon/awssdk/enhanced/dynamodb/DynamoDbTable.html) resource object.
The enhanced client's `table()` method creates a `DynamoDbTable` instance and requires parameters for the DynamoDB table name and a `DocumentTableSchema`.
The builder for a [DocumentTableSchema](https://sdk.amazonaws.com/java/api/latest/software/amazon/awssdk/enhanced/dynamodb/document/DocumentTableSchema.html) requires a primary index key and one or more attribute converter providers. The `AttributeConverterProvider.defaultProvider()` method provides converters for [default types](https://sdk.amazonaws.com/java/api/latest/software/amazon/awssdk/enhanced/dynamodb/internal/converter/attribute/package-summary.html). It should be specified even if you provide a custom attribute converter provider. You can add an optional secondary index key to the builder.
The following code snippet shows the code that generates the client-side representation of a DynamoDB `person` table that stores schemaless `EnhancedDocument` objects.
```
DynamoDbTable documentDynamoDbTable =
enhancedClient.table("person",
TableSchema.documentSchemaBuilder()
// Specify the primary key attributes.
.addIndexPartitionKey(TableMetadata.primaryIndexName(),"id", AttributeValueType.S)
.addIndexSortKey(TableMetadata.primaryIndexName(), "lastName", AttributeValueType.S)
// Specify attribute converter providers. Minimally add the default one.
.attributeConverterProviders(AttributeConverterProvider.defaultProvider())
.build());
// Call documentTable.createTable() if "person" does not exist in DynamoDB.
// createTable() should be called only one time.
```
The following shows the JSON representation of a `person` object that is used throughout this section.
### JSON `person` object
```
{
"id": 1,
"firstName": "Richard",
"lastName": "Roe",
"age": 25,
"addresses":
{
"home": {
"zipCode": "00000",
"city": "Any Town",
"state": "FL",
"street": "123 Any Street"
},
"work": {
"zipCode": "00001",
"city": "Anywhere",
"state": "FL",
"street": "100 Main Street"
}
},
"hobbies": [
"Hobby 1",
"Hobby 2"
],
"phoneNumbers": [
{
"type": "Home",
"number": "555-0100"
},
{
"type": "Work",
"number": "555-0119"
}
]
}
```
# Build enhanced documents
An `[EnhancedDocument](https://sdk.amazonaws.com/java/api/latest/software/amazon/awssdk/enhanced/dynamodb/document/EnhancedDocument.html)` represents a document-type object that has complex structure with nested attributes. An `EnhancedDocument` requires top-level attributes that match the primary key attributes specified for the `DocumentTableSchema`. The remaining content is arbitrary and can consist of top-level attributes and also deeply nested attributes.
You create an `EnhancedDocument` instance by using a builder that provides several ways to add elements.
## Build from a JSON string
With a JSON string, you can build an `EnhancedDocument` in one method call. The following snippet creates an `EnhancedDocument` from a JSON string returned by the `jsonPerson()` helper method. The `jsonPerson()` method returns the JSON string version of the [person object](ddb-en-client-doc-api-steps.md#ddb-en-client-doc-api-steps-createschema-obj) shown previously.
```
EnhancedDocument document =
EnhancedDocument.builder()
.json( jsonPerson() )
.build());
```
## Build from individual elements
Alternatively, you can build an `EnhancedDocument` instance from individual components using type-safe methods of the builder.
The following example builds a `person` enhanced document similar to the enhanced document that is built from the JSON string in the previous example.
```
/* Define the shape of an address map whose JSON representation looks like the following.
Use 'addressMapEnhancedType' in the following EnhancedDocument.builder() to simplify the code.
"home": {
"zipCode": "00000",
"city": "Any Town",
"state": "FL",
"street": "123 Any Street"
}*/
EnhancedType> addressMapEnhancedType =
EnhancedType.mapOf(EnhancedType.of(String.class), EnhancedType.of(String.class));
// Use the builder's typesafe methods to add elements to the enhanced document.
EnhancedDocument personDocument = EnhancedDocument.builder()
.putNumber("id", 50)
.putString("firstName", "Shirley")
.putString("lastName", "Rodriguez")
.putNumber("age", 53)
.putNull("nullAttribute")
.putJson("phoneNumbers", phoneNumbersJSONString())
/* Add the map of addresses whose JSON representation looks like the following.
{
"home": {
"zipCode": "00000",
"city": "Any Town",
"state": "FL",
"street": "123 Any Street"
}
} */
.putMap("addresses", getAddresses(), EnhancedType.of(String.class), addressMapEnhancedType)
.putList("hobbies", List.of("Theater", "Golf"), EnhancedType.of(String.class))
.build();
```
### Helper methods
```
private static String phoneNumbersJSONString() {
return " [" +
" {" +
" \"type\": \"Home\"," +
" \"number\": \"555-0140\"" +
" }," +
" {" +
" \"type\": \"Work\"," +
" \"number\": \"555-0155\"" +
" }" +
" ]";
}
private static Map> getAddresses() {
return Map.of(
"home", Map.of(
"zipCode", "00002",
"city", "Any Town",
"state", "ME",
"street", "123 Any Street"));
}
```
# Perform CRUD operations
After you define an `EnhancedDocument` instance, you can save it to a DynamoDB table. The following code snippet uses the [personDocument](ddb-en-client-doc-api-steps-create-ed.md#ddb-en-client-doc-api-steps-create-ed-fromparts) that was created from individual elements.
```
documentDynamoDbTable.putItem(personDocument);
```
After you read an enhanced document instance from DynamoDB, you can extract the individual attribute values using getters as shown in the following code snippet that access the data saved from the `personDocument`. Alternatively, you can extract the complete content to a JSON string as shown in the last part of the example code.
```
// Read the item.
EnhancedDocument personDocFromDb = documentDynamoDbTable.getItem(Key.builder().partitionValue(50).build());
// Access top-level attributes.
logger.info("Name: {} {}", personDocFromDb.getString("firstName"), personDocFromDb.getString("lastName"));
// Name: Shirley Rodriguez
// Typesafe access of a deeply nested attribute. The addressMapEnhancedType shown previously defines the shape of an addresses map.
Map> addresses = personDocFromDb.getMap("addresses", EnhancedType.of(String.class), addressMapEnhancedType);
addresses.keySet().forEach(k -> logger.info(addresses.get(k).toString()));
// {zipCode=00002, city=Any Town, street=123 Any Street, state=ME}
// Alternatively, work with AttributeValue types checking along the way for deeply nested attributes.
Map addressesMap = personDocFromDb.getMapOfUnknownType("addresses");
addressesMap.keySet().forEach((String k) -> {
logger.info("Looking at data for [{}] address", k);
// Looking at data for [home] address
AttributeValue value = addressesMap.get(k);
AttributeValue cityValue = value.m().get("city");
if (cityValue != null) {
logger.info(cityValue.s());
// Any Town
}
});
List phoneNumbers = personDocFromDb.getListOfUnknownType("phoneNumbers");
phoneNumbers.forEach((AttributeValue av) -> {
if (av.hasM()) {
AttributeValue type = av.m().get("type");
if (type.s() != null) {
logger.info("Type of phone: {}", type.s());
// Type of phone: Home
// Type of phone: Work
}
}
});
String jsonPerson = personDocFromDb.toJson();
logger.info(jsonPerson);
// {"firstName":"Shirley","lastName":"Rodriguez","addresses":{"home":{"zipCode":"00002","city":"Any Town","street":"123 Any Street","state":"ME"}},"hobbies":["Theater","Golf"],
// "id":50,"nullAttribute":null,"age":53,"phoneNumbers":[{"number":"555-0140","type":"Home"},{"number":"555-0155","type":"Work"}]}
```
`EnhancedDocument` instances can be used with any method of `[DynamoDbTable](https://sdk.amazonaws.com/java/api/latest/software/amazon/awssdk/enhanced/dynamodb/DynamoDbTable.html)` or [DynamoDbEnhancedClient](https://sdk.amazonaws.com/java/api/latest/software/amazon/awssdk/enhanced/dynamodb/DynamoDbEnhancedClient.html) in place of mapped data classes.
# Access enhanced document attributes as custom objects
In addition to providing an API to read and write attributes with schemaless structures, the Enhanced Document API lets you convert attributes to and from instances of custom classes.
The Enhanced Document API uses `AttributeConverterProvider`s and `AttributeConverter`s that were shown in the [control attribute conversion](ddb-en-client-adv-features-conversion.md) section as part of the DynamoDB Enhanced Client API.
In the following example, we use a `CustomAttributeConverterProvider` with its nested `AddressConverter` class to convert `Address` objects.
This example shows that you can mix data from classes and also data from structures that are built as needed. This example also shows that custom classes can be used at any level of a nested structure. The `Address` objects in this example are values used in a map.
```
public static void attributeToAddressClassMappingExample(DynamoDbEnhancedClient enhancedClient, DynamoDbClient standardClient) {
String tableName = "customer";
// Define the DynamoDbTable for an enhanced document.
// The schema builder provides methods for attribute converter providers and keys.
DynamoDbTable documentDynamoDbTable = enhancedClient.table(tableName,
DocumentTableSchema.builder()
// Add the CustomAttributeConverterProvider along with the default when you build the table schema.
.attributeConverterProviders(
List.of(
new CustomAttributeConverterProvider(),
AttributeConverterProvider.defaultProvider()))
.addIndexPartitionKey(TableMetadata.primaryIndexName(), "id", AttributeValueType.N)
.addIndexSortKey(TableMetadata.primaryIndexName(), "lastName", AttributeValueType.S)
.build());
// Create the DynamoDB table if needed.
documentDynamoDbTable.createTable();
waitForTableCreation(tableName, standardClient);
// The getAddressesForCustomMappingExample() helper method that provides 'addresses' shows the use of a custom Address class
// rather than using a Map to hold the address data.
Map addresses = getAddressesForCustomMappingExample();
// Build an EnhancedDocument instance to save an item with a mix of structures defined as needed and static classes.
EnhancedDocument personDocument = EnhancedDocument.builder()
.putNumber("id", 50)
.putString("firstName", "Shirley")
.putString("lastName", "Rodriguez")
.putNumber("age", 53)
.putNull("nullAttribute")
.putJson("phoneNumbers", phoneNumbersJSONString())
// Note the use of 'EnhancedType.of(Address.class)' instead of the more generic
// 'EnhancedType.mapOf(EnhancedType.of(String.class), EnhancedType.of(String.class))' that was used in a previous example.
.putMap("addresses", addresses, EnhancedType.of(String.class), EnhancedType.of(Address.class))
.putList("hobbies", List.of("Hobby 1", "Hobby 2"), EnhancedType.of(String.class))
.build();
// Save the item to DynamoDB.
documentDynamoDbTable.putItem(personDocument);
// Retrieve the item just saved.
EnhancedDocument srPerson = documentDynamoDbTable.getItem(Key.builder().partitionValue(50).sortValue("Rodriguez").build());
// Access the addresses attribute.
Map srAddresses = srPerson.get("addresses",
EnhancedType.mapOf(EnhancedType.of(String.class), EnhancedType.of(Address.class)));
srAddresses.keySet().forEach(k -> logger.info(addresses.get(k).toString()));
documentDynamoDbTable.deleteTable();
// The content logged to the console shows that the saved maps were converted to Address instances.
Address{street='123 Main Street', city='Any Town', state='NC', zipCode='00000'}
Address{street='100 Any Street', city='Any Town', state='NC', zipCode='00000'}
```
## `CustomAttributeConverterProvider` code
```
public class CustomAttributeConverterProvider implements AttributeConverterProvider {
private final Map, AttributeConverter> converterCache = ImmutableMap.of(
// 1. Add AddressConverter to the internal cache.
EnhancedType.of(Address.class), new AddressConverter());
public static CustomAttributeConverterProvider create() {
return new CustomAttributeConverterProvider();
}
// 2. The enhanced client queries the provider for attribute converters if it
// encounters a type that it does not know how to convert.
@SuppressWarnings("unchecked")
@Override
public AttributeConverter converterFor(EnhancedType enhancedType) {
return (AttributeConverter) converterCache.get(enhancedType);
}
// 3. Custom attribute converter
private class AddressConverter implements AttributeConverter {
// 4. Transform an Address object into a DynamoDB map.
@Override
public AttributeValue transformFrom(Address address) {
Map attributeValueMap = Map.of(
"street", AttributeValue.fromS(address.getStreet()),
"city", AttributeValue.fromS(address.getCity()),
"state", AttributeValue.fromS(address.getState()),
"zipCode", AttributeValue.fromS(address.getZipCode()));
return AttributeValue.fromM(attributeValueMap);
}
// 5. Transform the DynamoDB map attribute to an Address oject.
@Override
public Address transformTo(AttributeValue attributeValue) {
Map m = attributeValue.m();
Address address = new Address();
address.setStreet(m.get("street").s());
address.setCity(m.get("city").s());
address.setState(m.get("state").s());
address.setZipCode(m.get("zipCode").s());
return address;
}
@Override
public EnhancedType type() {
return EnhancedType.of(Address.class);
}
@Override
public AttributeValueType attributeValueType() {
return AttributeValueType.M;
}
}
}
```
## `Address` class
```
public class Address {
private String street;
private String city;
private String state;
private String zipCode;
public Address() {
}
public String getStreet() {
return this.street;
}
public String getCity() {
return this.city;
}
public String getState() {
return this.state;
}
public String getZipCode() {
return this.zipCode;
}
public void setStreet(String street) {
this.street = street;
}
public void setCity(String city) {
this.city = city;
}
public void setState(String state) {
this.state = state;
}
public void setZipCode(String zipCode) {
this.zipCode = zipCode;
}
}
```
## Helper method that provides addresses
The following helper method provides the map that use custom `Address` instances for values rather than generic `Map` instances for values.
```
private static Map getAddressesForCustomMappingExample() {
Address homeAddress = new Address();
homeAddress.setStreet("100 Any Street");
homeAddress.setCity("Any Town");
homeAddress.setState("NC");
homeAddress.setZipCode("00000");
Address workAddress = new Address();
workAddress.setStreet("123 Main Street");
workAddress.setCity("Any Town");
workAddress.setState("NC");
workAddress.setZipCode("00000");
return Map.of("home", homeAddress,
"work", workAddress);
}
```
# Use an `EnhancedDocument` without DynamoDB
Although you usually use an instance of an `EnhancedDocument` to read and write document-type DynamoDB items, it can also be used independently of DynamoDB.
You can use `EnhancedDocuments` for their ability to convert between JSON strings or custom objects to low-level maps of `AttributeValues` as shown in the following example.
```
public static void conversionWithoutDynamoDbExample() {
Address address = new Address();
address.setCity("my city");
address.setState("my state");
address.setStreet("my street");
address.setZipCode("00000");
// Build an EnhancedDocument instance for its conversion functionality alone.
EnhancedDocument addressEnhancedDoc = EnhancedDocument.builder()
// Important: You must specify attribute converter providers when you build an EnhancedDocument instance not used with a DynamoDB table.
.attributeConverterProviders(new CustomAttributeConverterProvider(), DefaultAttributeConverterProvider.create())
.put("addressDoc", address, Address.class)
.build();
// Convert address to a low-level item representation.
final Map addressAsAttributeMap = addressEnhancedDoc.getMapOfUnknownType("addressDoc");
logger.info("addressAsAttributeMap: {}", addressAsAttributeMap.toString());
// Convert address to a JSON string.
String addressAsJsonString = addressEnhancedDoc.getJson("addressDoc");
logger.info("addressAsJsonString: {}", addressAsJsonString);
// Convert addressEnhancedDoc back to an Address instance.
Address addressConverted = addressEnhancedDoc.get("addressDoc", Address.class);
logger.info("addressConverted: {}", addressConverted.toString());
}
/* Console output:
addressAsAttributeMap: {zipCode=AttributeValue(S=00000), state=AttributeValue(S=my state), street=AttributeValue(S=my street), city=AttributeValue(S=my city)}
addressAsJsonString: {"zipCode":"00000","state":"my state","street":"my street","city":"my city"}
addressConverted: Address{street='my street', city='my city', state='my state', zipCode='00000'}
*/
```
**Note**
When you use an enhanced document independent of a DynamoDB table, make sure you explicitly set attribute converter providers on the builder.
In contrast, the document table schema supplies the converter providers when an enhanced document is used with a DynamoDB table.