Dave Lang of the Amazon DynamoDB team is back with another guest post!
In January we launched Amazon DynamoDB, our high scale, low latency, fully-managed NoSQL database service. Since our launch, one of the most frequent requests we’ve heard from customers is for a way to write multiple items in a single request to increase parallelism and speed up the process of loading data into DynamoDB.
I’m happy to let you know that we’ve released that feature today. You can now use DynamoDB’s new BatchWriteItem feature to add, delete, or replace up to 25 items at a time.
BatchWriteItem makes it easier to load large amounts of data into DynamoDB. By combining multiple writes in a single request, BatchWriteItem allows you to achieve parallelism without having to manage it yourself at the application level. You should certainly consider using BatchWriteItem if you are working in a language that makes it harder to write multi-threaded applications.
Since BatchWriteItem will add/delete/replace items in parallel, you can also expect better average latencies than when you use individual PutItem or DeleteItem requests. When used to add multiple items, BatchWriteItem is designed to provide average latencies that are similar to the latency of a single PutItem operation.
I’d like to thank our customers for letting us know (loud and clear!) that you wanted this feature. You posted to our forum, sent us emails, and sent comments during our webinars. This kind of feedback helps us build the features you need, so please keep telling us what you’d like to see next.
To learn more about how to use BatchWriteItem, you can read the DynamoDB Documentation.
Make sure you have the latest SDK libraries so that you can take advantage of this new feature. You can download them at http://aws.amazon.com/code.
- Dave Lang