CODEWITHSUNDEEP
javadatabaseamazon-dynamodb
Ali
Ali

Reputation: 267059

Do global secondary indexes in DynamoDB need to be unique?

Say I have an order table which stores orders made by users. It has the following schema:

id (string, unique primary key)

userId (string, id of the user who made the user

I'd like to add a global secondary index on the userId column, but it will be non-unique, i.e a single user may make multiple orders, which means there might be multiple orders with the same userId.

From reading the docs, I can't quite tell if this scenario would be ok, or if I need to add a range key to make the primary key be unique.

I'd prefer not to have to add a range key, so I can simply use the userId to get all the orders made by a particular user.

Any thoughts?

Upvotes: 3

Views: 2584

Answers (1)

Eyal Ch
Eyal Ch

Reputation: 10056

this scenario would be ok, global secondary index can be non unique. the only disadvantage (that i can think about) is that if you have a 'mega' user, with multiple orders (on write), it can accuse throughput capacity errors..

a good reference is here:

http://docs.aws.amazon.com/amazondynamodb/latest/developerguide/GuidelinesForTables.html

Upvotes: 2

Related Questions