Okay, so this is *really* handy for managing composite secondary indexes on DynamoDB tables. Way too often we run into situations where we want to upsert a record, but we have incomplete data to regenerate the composite key. For example, let's say we have a composite sort key for "{country}#{state}#{city}". If we want to update just the city, we need to have the country and state attributes available so we can properly rebuild the composite key. With this update, we can now update just the ONE attribute and the keys are automatically regenerated for us! Super useful when you don't have the entire object when upserting (like partial data from events). Also, sorting respects each attribute's type (string, number, or binary), so composite sort keys that are mixed like "{customerName}#{invoiceNo}" no longer need to be stored as "acme#02031" and "acme#00045" to lexicographically sort correctly. Now you essentially get accurate sort order with the equivalent of "acme#2031" and "acme#45". No more zero-padding! Oh yeah, you also get AUTOMATIC BACKFILLING with existing attributes. Create a new index using existing attributes and you magically have all your composite keys generated for you! I haven't spent a lot of time with this, but the query interface will be a bit more tricky as you'll need to specify all the matching attributes in your key condition expression. But who knows, maybe someone will build an open source library to make it easier. 😀 aws.amazon.com/blogs/databas…