How to set DynamoDB range key, String or Map

0
votes

I have a DynamoDB table with a primary hash key, and a range key. Range key will have two attributes. Say those attribute names are: name1, name2, with values value1, value2

Plan A: combine two attributes as string, use comma as delimiter
Primary hash key: id
Range key: value1,value2

Cons
1. comma may not work if some wired values contain this delimiter

Plan B: convert map as String for range key
Primary hash key: id
Range key: “{\“name1\”: \“value1\”, \“name2\”: \“value2\”}”

Cons
1. different SDK may result into different JSON String based on the same value? (Not sure), need to support multiple SDK read/write. Like Java and Ruby

So, which solution works better? Or there are any better suggestions?

Thanks! Ray

1
amazon-dynamodb

1 Answers

0
votes

You're on the right track. The AWS docs regarding key design promote your first suggestion, but it also has some warnings about the situation that you refered as cons.

I don't think that you could have problemas with different sdk parsers, but I also think that a little bit of precautions here would be a good ideia. So instead of directly parse a json to string using the sdk, I would manually concatenate the values using a custom function to generate a deterministic value like "name1-value1-name2-value2" or "name1:value1-name2:value2".