BatchWriteItem

Description

This operation enables you to put or delete several items across multiple tables in a single call.

To upload one item, you can use PutItem, and to delete one item, you can use DeleteItem. However, when you want to upload or delete large amounts of data, such as uploading large amounts of data from Amazon EMR (Amazon EMR) or migrating data from another database in to DynamoDB, BatchWriteItem offers an efficient alternative.

If you use languages such as Java, you can use threads to upload items in parallel. This adds complexity in your application to handle the threads. Other languages don't support threading. For example, if you are using PHP, you must upload or delete items one at a time. In both situations, BatchWriteItem provides an alternative where the specified put and delete operations are processed in parallel, giving you the power of the thread pool approach without having to introduce complexity in your application.

Note that each individual put and delete specified in a BatchWriteItem operation costs the same in terms of consumed capacity units. However, because BatchWriteItem performs the specified operations in parallel, you get lower latency. Delete operations on non-existent items consume 1 write capacity unit. For more information about consumed capacity units, see Working with tables and data in DynamoDB.

When using BatchWriteItem, note the following limitations:

DynamoDB rejects the entire batch write operation if any one of the following is true:

Requests

Syntax

// This header is abbreviated. For a sample of a complete header, see DynamoDB low-level API.
POST / HTTP/1.1
x-amz-target: DynamoDB_20111205.BatchGetItem 
content-type: application/x-amz-json-1.0 

{
  "RequestItems" :  RequestItems
} 

RequestItems
{
     "TableName1" :  [ Request, Request, ... ],
     "TableName2" :  [ Request, Request, ... ],
     ...
}

Request ::=
  PutRequest | DeleteRequest

PutRequest ::=
{
  "PutRequest" : {
     "Item" : {
        "Attribute-Name1" : Attribute-Value,
        "Attribute-Name2" : Attribute-Value,
         ...
     }
  }
}

DeleteRequest ::=
{
   "DeleteRequest" : {
      "Key" : PrimaryKey-Value
   }
}


PrimaryKey-Value ::= HashTypePK | HashAndRangeTypePK

HashTypePK ::=
{ 
   "HashKeyElement" : Attribute-Value
}

HashAndRangeTypePK
{ 
   "HashKeyElement" : Attribute-Value,
   "RangeKeyElement" : Attribute-Value, 
}

Attribute-Value ::= String | Numeric| Binary | StringSet | NumericSet | BinarySet 

Numeric ::=
{
   "N": "Number"
}

String ::=
{
   "S": "String"
}

Binary ::=
{
    "B": "Base64 encoded binary data"
}

StringSet ::=
{
   "SS": [ "String1", "String2", ... ]
}

NumberSet ::=
{
   "NS": [ "Number1", "Number2", ... ]
}  

BinarySet ::=
{
   "BS": [ "Binary1", "Binary2", ... ]
}

In the request body, the RequestItems JSON object describes the operations that you want to perform. The operations are grouped by tables. You can use BatchWriteItem to update or delete several items across multiple tables. For each specific write request, you must identify the type of request (PutItem, DeleteItem) followed by detail information about the operation.

Responses

Syntax

The following is the syntax of the JSON body returned in the response.

{
  "Responses" :         ConsumedCapacityUnitsByTable
  "UnprocessedItems" :  RequestItems
} 

ConsumedCapacityUnitsByTable
{
    "TableName1" : { "ConsumedCapacityUnits", : NumericValue },
    "TableName2" : { "ConsumedCapacityUnits", : NumericValue },
     ...
}

RequestItems
This syntax is identical to the one described in the JSON syntax in the request.

Special errors

No errors specific to this operation.

Examples

The following example shows an HTTP POST request and the response of a BatchWriteItem operation. The request specifies the following operations on the Reply and the Thread tables:

For examples using the AWS SDK, see Working with items and attributes in DynamoDB.

Sample request

// This header is abbreviated. For a sample of a complete header, see DynamoDB low-level API.
POST / HTTP/1.1
x-amz-target: DynamoDB_20111205.BatchGetItem 
content-type: application/x-amz-json-1.0 

{
  "RequestItems":{
    "Reply":[
      {
        "PutRequest":{
          "Item":{
            "ReplyDateTime":{
              "S":"2012-04-03T11:04:47.034Z"
            },
            "Id":{
              "S":"DynamoDB#DynamoDB Thread 5"
            }
          }
        }
      },
      {
        "DeleteRequest":{
          "Key":{
            "HashKeyElement":{
              "S":"DynamoDB#DynamoDB Thread 4"
            },
            "RangeKeyElement":{
              "S":"oops - accidental row"
            }
          }
        }
      }
    ],
    "Thread":[
      {
        "PutRequest":{
          "Item":{
            "ForumName":{
              "S":"DynamoDB"
            },
            "Subject":{
              "S":"DynamoDB Thread 5"
            }
          }
        }
      }
    ]
  }
}

Sample response

The following example response shows a put operation on both the Thread and Reply tables succeeded and a delete operation on the Reply table failed (for reasons such as throttling that is caused when you exceed the provisioned throughput on the table). Note the following in the JSON response:

HTTP/1.1 200 OK
x-amzn-RequestId: G8M9ANLOE5QA26AEUHJKJE0ASBVV4KQNSO5AEMVJF66Q9ASUAAJG
Content-Type: application/x-amz-json-1.0
Content-Length: 536
Date: Thu, 05 Apr 2012 18:22:09 GMT

{
   "Responses":{
      "Thread":{
         "ConsumedCapacityUnits":1.0
      },
      "Reply":{
         "ConsumedCapacityUnits":1.0
      }
   },
   "UnprocessedItems":{
      "Reply":[
         {
            "DeleteRequest":{
               "Key":{
                  "HashKeyElement":{
                     "S":"DynamoDB#DynamoDB Thread 4"
                  },
                  "RangeKeyElement":{
                     "S":"oops - accidental row"
                  }
               }
            }
         }
      ]
   }
}