使用 AWS Amplify 更新 DynamoDB 架构

Question

According to the AWS Amplify documentation:

• " objects annotated with @model are stored in Amazon DynamoDB ";
• " a single @model directive configures... an Amazon DynamoDB table "; and
• one can " push updated changes with amplify push ".

It seems clear that amplify push creates a DynamoDB table for each @model.

My questions relate to schema updates:

• I imagine that adding/removing a model or adding/removing a field in a model works by updating the schema document and then running amplify push . Is that right?

• How does one rename a model or a field? How would amplify push know to rename vs. drop the old and add the new?

• How does one implement a migration that requires some business logic, eg, to update the contents of existing rows? Doing this without Amplify has already been addressed but it is unclear whether that would conflict with something that amplify push might try to do.

Answer 1

DynamoDB is schema-less, and doesn't care about your application schema as long as you don't try to change its hash key or range key

Therefore, nothing really happens on the datastore side. If you drop a key and add a new one in your schema, then your application will start to search and write data to the new key. Old key will simply be ignored from now on, but existing data will be kept in the datastore.

If you want to rename a key, then you would have to migrate the data by yourself through mass update on the table. There are many ways to do it, the simpliest one being scaning the table and performing updates on found items.

Answer 2

你有没有试过用这个编译模式：

amplify api gql-compile

Answer 3

Have you tried

amplify codegen models

before doing the

amplify push

after?