Custom DynamoDb TableNameResolver not being called when using CrudRepository - spring-data-jpa

Question

I am testing DynamoDB tables and want to set up different table names for prod and dev environment using the keyword"dev" for development and prod for production.
I have a POJO
#DynamoDBTable(tableName = "abc_xy_dev_MyProjectName_Employee")
public class Employee implements Cloneable {
}
On Prod I want its name to be abc_xy_prod_MyProjectName_Employee.
So, I wrote a TableNameResolver
public static class MyTableNameResolver implements TableNameResolver {
public static final MyTableNameResolver INSTANCE = new MyTableNameResolver();
#Override
public String getTableName(Class<?> clazz, DynamoDBMapperConfig config) {
final TableNameOverride override = config.getTableNameOverride();
String tableNameToReturn = null;
if (override != null) {
final String tableName = override.getTableName();
if (tableName != null) {
System.out.println("MyTableNameResolver ==================================");
return tableName;
}
}
String env = System.getenv("DEPLOYMENT_ENV");
for(Annotation annotation : clazz.getAnnotations()){
if(annotation instanceof DynamoDBTable){
DynamoDBTable myAnnotation = (DynamoDBTable) annotation;
if ("production".equals(env)){
tableNameToReturn = myAnnotation.tableName().replace("_dev_", "_prod_");
}
else {
tableNameToReturn = myAnnotation.tableName();
}
}
}
return tableNameToReturn;
}
}
This works by creating a table with the name abc_xy_prod_MyProjectName_Employee in production.
However, I have a repository with the following code
#EnableScan
public interface EmployeeRepository extends CrudRepository<Employee, String>
{
#Override
<S extends Employee> S save(S employee);
Optional<Employee> findById(String id);
#Override
List<Employee> findAll();
Optional<Employee> findByEmployeeNumber(String EmployeeNumber);
}
Thus when i try to call the method findAll via a endpoint /test case, i get the exception
There was an unexpected error (type=Internal Server Error,
status=500). User:
arn:aws:iam::87668976786:user/svc_nac_ps_MyProjectName_prod is not
authorized to perform: dynamodb:Scan on resource:
:table/abc_xy_dev_MyProjectName_Employee (Service: AmazonDynamoDBv2;
Status Code: 400; Error Code: AccessDeniedException; Request ID:
aksdnhLDFL)
i.e MyTableNameResolver doesn't get called internally when the respository methods are executed. It still points to table name with the name abc_xy_dev_MyProjectName_Employee given in the annotation #DynamoDBTable(tableName = "abc_xy_dev_MyProjectName_Employee")

Answer 1

You have used spring JPA as persistence dynamoDB Integration.
Below configuration can be used to set table name override as part of spring boot configuration.
Sample example is found in https://github.com/ganesara/SpringExamples/tree/master/spring-dynamo
Map Dynamo db repository with user defined mapper config reference
#EnableDynamoDBRepositories(basePackages = "home.poc.spring", dynamoDBMapperConfigRef="dynamoDBMapperConfig")
Mapper Config for table override is as below
#Bean
public DynamoDBMapperConfig dynamoDBMapperConfig() {
DynamoDBMapperConfig mapperConfig = new DynamoDBMapperConfig
.Builder()
.withTableNameOverride(DynamoDBMapperConfig.TableNameOverride.withTableNamePrefix("PROD_"))
.build();
return mapperConfig;
}
Full configuration for reference
#Configuration
#EnableDynamoDBRepositories(basePackages = "home.poc.spring", dynamoDBMapperConfigRef="dynamoDBMapperConfig")
public class DynamoDBConfig {
#Value("${amazon.dynamodb.endpoint}")
private String amazonDynamoDBEndpoint;
#Value("${amazon.aws.accesskey}")
private String amazonAWSAccessKey;
#Value("${amazon.aws.secretkey}")
private String amazonAWSSecretKey;
#Bean
public AmazonDynamoDB amazonDynamoDB() {
AmazonDynamoDB amazonDynamoDB
= new AmazonDynamoDBClient(amazonAWSCredentials());
if (!StringUtils.isEmpty(amazonDynamoDBEndpoint)) {
amazonDynamoDB.setEndpoint(amazonDynamoDBEndpoint);
}
return amazonDynamoDB;
}
#Bean
public AWSCredentials amazonAWSCredentials() {
return new BasicAWSCredentials(
amazonAWSAccessKey, amazonAWSSecretKey);
}
#Bean
public DynamoDBMapperConfig dynamoDBMapperConfig() {
DynamoDBMapperConfig mapperConfig = new DynamoDBMapperConfig
.Builder()
.withTableNameOverride(DynamoDBMapperConfig.TableNameOverride.withTableNamePrefix("PROD_"))
.build();
return mapperConfig;
}
#Bean
public DynamoDBMapper dynamoDBMapper() {
return new DynamoDBMapper(amazonDynamoDB(), dynamoDBMapperConfig());
}
}

Answer 2

You are using DynamoDBMapper (the Java SDK). Here is how I use it. Lets say I have a table called Users, with an associated User POJO. In DynamoDB I have DEV_Users and LIVE_Users.
I have an environment variable 'ApplicationEnvironmentName' which is either DEV or LIVE.
I create a custom DynamoDBMapper like this:
public class ApplicationDynamoMapper {
private static Map<String, DynamoDBMapper> mappers = new HashMap<>();
private static AmazonDynamoDB client = AmazonDynamoDBClientBuilder.standard()
.withRegion(System.getProperty("DynamoDbRegion")).build();
protected ApplicationDynamoMapper() {
// Exists only to defeat instantiation.
}
public static DynamoDBMapper getInstance(final String tableName) {
final ApplicationLogContext LOG = new ApplicationLogContext();
DynamoDBMapper mapper = mappers.get(tableName);
if (mapper == null) {
final String tableNameOverride = System.getProperty("ApplicationEnvironmentName") + "_" + tableName;
LOG.debug("Creating DynamoDBMapper with overridden tablename {}.", tableNameOverride);
final DynamoDBMapperConfig mapperConfig = new DynamoDBMapperConfig.Builder().withTableNameOverride(TableNameOverride.withTableNameReplacement(tableNameOverride)).build();
mapper = new DynamoDBMapper(client, mapperConfig);
mappers.put(tableName, mapper);
}
return mapper;
}
}
My Users POJO looks like this:
#DynamoDBTable(tableName = "Users")
public class User {
...
}
When I want to use the database I create an application mapper like this:
DynamoDBMapper userMapper = ApplicationDynamoMapper.getInstance(User.DB_TABLE_NAME);
If I wanted to a load a User, I would do it like this:
User user = userMapper.load(User.class, userId);
Hope that helps.