I'm new to DynamoDB. Following the best practice of DynamoDB, I'm going to create only one single table, but wonder how the partition and sort keys should be named.
In my understanding, the keys should have abstract names such as as "PK" and "SK", because the single table stores versatile data in the same key. Is this right?
Or, some content-specific names should be given?
Primary keys are easiest to use if they are named {table}Id. For example customerId, orderId etc.
Secondary keys, meaning indexes, can have their original name, as you can set an index on any column at any point in time, so no need to name in a specific way.
For example (User table): userId, name, companyId, email... Where companyId is a secondary index.
You are correct, "PK" and "SK", or lowercase "pk" and "sk" is very common and useful because tables will store a mixture of entity types. Similarly, GSI keys are often named "gsi1pk" and "gsi1sk". Another good reason, aside from convention, is to save on storage. Each item in DynamoDB stores both the key and value for every attribute. Long attribute names can consume a large portion of your storage.
