what is the most efficient way to update bulk data inside DynamoDB rather than scan and query

I've got a table containing loads of category keys. I want to update some of these categories by a bulk request(e.g. update key#1, key#3 ) and the table schema is as follows:

as you can see there is no key#3 inside the table. I want to write an efficient query to update ( key#1, key#3 ), according to my request which is a bulk one.
so I want the most efficient way for doing this and simultaneously I don't want my updates to fail without being noticed.

• I know one way would be doing a scan operation to see if the table contains my array, which is not efficient at all.
• another one would be using a loop to query over the database to mark which keys are present and which are not. Then try to update the remaining ones. I don't think this might be a good idea either.
• The last thing that came into my mind and I think might work, would be trying to update the database using transact Items containing all the updates without checking if any of those records exist. cause I hope this one returns an error if the update operation fails for any of those keys. and intuitively I think this might be a more efficient way ( It comes from my SQL background).

so I want to know if there are any more efficient ways rather than the above three?
and if my latest approach would be a good one or not.
and what are the pros and cons of doing that?

After lots of digging, I found no straight solutions for this one. But we can somehow simulate it using conditional expressions and utilizing native programming language's features like multithreading, multiprocessing, and similar.

Here are what I've found:

• The way to achieve this is by the means of conditional expressions in which you can define the status of an attribute or its existence status.
• The problem is the conditional expressions are not applicable in batchWriteItems, and they are only applicable inside TransactWriteItems.
• TransactWriteItems are two times slower than normal operations. So regarding performance requirements, this won't be applicable.
• So the best way is to use single putItem with conditional expressions multiple times.
• To achieve the best performance we can leverage multithreading or concurrency or multiprocessing features according to the programming language we are using and call for example 100 queries simultaneously in parallel (AWS-DynamoDB can handle that. just make sure you've taken care of throughput capacity).