Update attributes using UpdateItemEnhancedRequest - DynamoDbEnhancedClient - amazon-web-services

Question

As per DynamoDb Java Sdk2, the Update operation can be performed as
DynamoDbTable<Customer> mappedTable = enhancedClient.table("Customer", TableSchema.fromBean(Customer.class));
Key key = Key.builder()
.partitionValue(keyVal)
.build();
Customer customerRec = mappedTable.getItem(r->r.key(key));
customerRec.setEmail(email);
mappedTable.updateItem(customerRec);
Shouldn't this will cause two calls to dynamoDB
What if after fetching the record and before updateItem call, another thread updated the record, so we have to put it into a transaction as well
Although there is another way by using UpdateItemEnhancedRequest
final var request = UpdateItemEnhancedRequest.builder(Customer.class)
.item(updatedCustomerObj)
.ignoreNulls(Boolean.TRUE)
.build();
mappedTable.updateItem(request);
but this would require using ignoreNulls(TRUE) and will not handle updates in case null value is to be set.
What should be the optimized way for update operation using enhanced client

Answer 1

For 1:
I don't like this part of the SDK but the "non-enhanced" client can update without first querying for the entity.
UpdateItemRequest updateItemRequest = UpdateItemRequest
.builder()
.tableName("Customer")
.key(
Map.of(
"PK", AttributeValue.builder().s("pk").build(),
"SK", AttributeValue.builder().s("sk").build()
))
.attributeUpdates(
Map.of(
"email", AttributeValueUpdate.builder()
.action(AttributeAction.PUT)
.value(AttributeValue.builder().s("new-email#gmail.com").build())
.build()
))
.build();
client.updateItem(updateItemRequest);
For 2:
Yes, you are absolutely right that can happen and will lead to data inconsistency. To avoid this you should use conditional write/update with a version. Luckily we have a Java annotation for this
#DynamoDBVersionAttribute
public Long getVersion() { return version; }
public void setVersion(Long version) { this.version = version;}
More info here