Nova ODM / Amazon DynamoDB Batch Iteration

This library provides utilities for automatically submitting arbitrarily-sized batches of reads and writes to DynamoDB using well-formed BatchGetItem and BatchWriteItem operations, respectively. Partial successes (i.e., BatchGetItem operations that return some responses and some unprocessed keys or BatchWriteItem operations that return some unprocessed items) will retry the unprocessed items automatically using exponential backoff.

Getting started

Reading batches of items

Create a BatchGet object, supplying an instantiated DynamoDB client from the AWS SDK for JavaScript and an iterable of keys that you wish to retrieve. The iterable may be synchronous (such as an array) or asynchronous (such as an object stream wrapped with async-iter-stream's wrap method).

import { BatchGet } from '@nova-odm/batch-iterator';
import { DynamoDBClient } from '@aws-sdk/client-dynamodb';

const dynamoDb = new DynamoDBClient({region: 'us-west-2'});
const keys = [
    ['tableName', {keyProperty: {N: '0'}}],
    ['tableName', {keyProperty: {N: '1'}}],
    ['tableName', {keyProperty: {N: '2'}}],
    // etc., continuing to count up to
    ['tableName', {keyProperty: {N: '1001'}}],
];

for await (const item of new BatchGet(dynamoDb, keys)) {
    console.log(item);
}

The above code snippet will automatically split the provided keys into BatchGetItem requests of 100 or fewer keys, and any unprocessed keys will be automatically retried until they are handled. The above code will execute at least 11 BatchGetItem operations, dependening on how many items are returned without processing due to insufficient provisioned read capacity.

Each item yielded in the for...await...of loop will be a single DynamoDB record. Iteration will stop once each key has been retrieved or an error has been encountered.

Writing batches of items

Create a BatchWrite object, supplying an instantiated DynamoDB client from the AWS SDK for JavaScript and an iterable of write requests that you wish to execute. The iterable may be synchronous (such as an array) or asynchronous (such as an object stream wrapped with async-iter-stream's wrap method).

Each write request should contain either a DeleteRequest key or a PutRequest key as described in the Amazon DynamoDB API reference.

import { BatchWrite } from '@nova-odm/batch-iterator';
import { DynamoDBClient } from '@aws-sdk/client-dynamodb';

const dynamoDb = new DynamoDBClient({region: 'us-west-2'});
const keys = [
    ['tableName', {DeleteRequest: {Key: {keyProperty: {N: '0'}}}}],
    ['tableName', {PutRequest: {Item: {keyProperty: {N: '1'}, otherProperty: {BOOL: false}}}}],
    ['tableName', {DeleteRequest: {Key: {keyProperty: {N: '2'}}}}],
    ['tableName', {PutRequest: {Item: {keyProperty: {N: '3'}, otherProperty: {BOOL: false}}}}],
    ['tableName', {N: '2'}],
    // etc., continuing to count up to
    ['tableName', {DeleteRequest: {Key: {keyProperty: {N: '102'}}}}],
];

for await (const item of new BatchWrite(dynamoDb, keys)) {
    console.log(item);
}

The above code snippet will automatically split the provided keys into BatchWriteItem requests of 25 or fewer write request objects, and any unprocessed request objects will be automatically retried until they are handled. The above code will execute at least 5 BatchWriteItem operations, dependening on how many items are returned without processing due to insufficient provisioned write capacity.

Each item yielded in the for...await...of loop will be a single write request that has succeeded. Iteration will stop once each request has been handled or an error has been encountered.