Auto increment number (non primary key) with AppSync and DynamoDB

** Which Category is your question related to? ** API

** What AWS Services are you utilizing? ** AppSync, DynamoDB

** Provide additional details e.g. code snippets ** Let’s say I have below schema:

type Item @model {
  category: String!
  name: String!
  tagNumber: String!
}

What I want to implement:

id	category	name	tagNumber
3c174388-054b-431f-a0f4-d835acdd273a	food	cheese cake	1.1
63b189d0-597e-47f3-8dcf-996da87101a8	book	harry porter	2.1
87749278-acae-4de1-b0ff-84d6b86506cc	food	steak	1.2
fd2334a9-ff76-4ae2-9ca5-834d998f91e7	book	bad blood	2.2

If I want to add a new item, which is a book called Joke, it should have a tagNumber 2.3, because it belongs to the 2nd category in the table, and it is the 3rd item in that category.

Why do I want to implement this? Because the tagNumber is easiest for people, not computers, to look up items.

Is there a way to implement this with AppSync and DynamoDB? I can think of the hard way to do it: use pipeline resolver. If the category specified exists, find an item with the latest createdAt timestamp in that category with tagNumber x.y, and assign x.<y+1> to the new item. If the category specified doesn’t exist, count categories N, and assign <N+1>.1 to the new item.

[dabit3]

Coming back to this, if anyone wants a working example of an atomic counter (increment, decrement) with DynamoDB and AppSync, here’s something I’m using:

{
    "version": "2018-05-29",
    "operation": "UpdateItem",
    "key" : {
      "id" : $util.dynamodb.toDynamoDBJson($context.arguments.id)
    },
    "update": {
      "expression" : "set #upvotes = #upvotes + :updateValue",
      "expressionNames" : {
           "#upvotes" : "upvotes"
       },
       "expressionValues" : {
           ":updateValue" : { "N" : 1 }
       }
    }
}

[mikeparisstuff]

@YikSanChan We are working on designs that will make working with pipeline resolvers a lot easier. Please see aws-amplify/amplify-category-api#430 and offer feedback. It is going to go through another round of review and design and it would be great to cover this use case.