feat(connector): add DynamoDB sink

[jetjinser]

I hereby agree to the terms of the RisingWave Labs, Inc. Contributor License Agreement.

What's changed and what's your intention?

1:1 mapping DynamoDB table sink.

Checklist

• I have written necessary rustdoc comments
• I have added necessary unit tests and integration tests
• I have added test labels as necessary. See details.
• I have added fuzzing tests or opened an issue to track them. (Optional, recommended for new SQL features Sqlsmith: Sql feature generation #7934).
• My PR contains breaking changes. (If it deprecates some features, please create a tracking issue to remove them in the future).
• All checks passed in ./risedev check (or alias, ./risedev c)
• My PR changes performance-critical code. (Please run macro/micro-benchmarks and show the results.)

• My PR contains critical fixes that are necessary to be merged into the latest release. (Please check out the details)

Documentation

• My PR needs documentation updates. (Please use the Release note section below to summarize the impact on users)

Release note

If this PR includes changes that directly affect users or other significant modifications relevant to the community, kindly draft a release note to provide a concise summary of these changes. Please prioritize highlighting the impact these changes will have on users.

new sink:

CREATE TABLE IF NOT EXISTS movies (
    year integer,
    title varchar,
    description varchar,
    primary key (year, title)
);

CREATE SINK
  dyn_sink
FROM
  movies
WITH
(
  connector = 'dynamodb',
  table = 'Movies',
  primary_key = 'year,title',
  endpoint = 'http://localhost:8000',
  region = 'us',
  access_key = 'ac',
  secret_key = 'sk'
);

[xiangjinwu]

• struct may be better mapped to DynamoDB Map
• jsonb can be mapped as string for now, and latter extended to a dynamic recursive one similar to #11699. It is your call whether the latter dynamic one shall be the default here instead. Just bringing the alternative option to attention.

[jetjinser]

• I'm not sure, the struct here must be named, since in create sink?
• Does this mean providing an TimestampHandlingMode-like option for jsonb format to DynamoDB sink?

[neverchanje]

You can consider struct as a list of named key values pairs, right?

struct <
  id varchar,
  name varchar
>

Map<String, Value> {
  "id": "123",
  "name": "jinser"
}

[jetjinser]

Yes, I think the current implementation is like this

[fuyufjh]

It seems that matching the data_type only could make code a bit cleaner

[jetjinser]

I'm worried that the DataType is different from the ScalarRefImpl variant, and just matching data_type would mean requiring scalar_ref.into_foo() (which could panic in some case). Do you mean that this function map_data_type is only used here and does not need to consider possible the tow varant different situations?

[fuyufjh]

I think each DataType must have only one possible corresponding ScalarRefImpl, so no need to match ScalarRefImpl.

[fuyufjh]

Have you ever tested an update event?

In self.payload_writer.write_chunk(), you inserted all the events before delete, which effectively means the UpdateInsert will happen before UpdateDelete. IIUC, the rows will be deleted rather than being updated.

Here, I prefer to handle UpdateInsert and UpdateDelete as a whole i.e. convert them into a single put KV into DynamoDB. There is no reason not doing so.

But keep in mind that this won't really solve the problem. Please carefully consider both of the following case:

• An Insert comes after Delete on the same key
• An Delete comes after Insert on the same key

Perhaps the current implementation of self.payload_writer.write_chunk() won't work.

[yuhao-su]

You can just put AwsAuthProps here.

[lmatz]

DynamoDB can be deployed locally via docker image: https://docs.aws.amazon.com/amazondynamodb/latest/developerguide/DynamoDBLocal.html
worth adding an integration test

[jetjinser]

DynamoDB can be deployed locally via docker image: https://docs.aws.amazon.com/amazondynamodb/latest/developerguide/DynamoDBLocal.html worth adding an integration test

Should I add it to this PR? Or should I open another PR?

[jetjinser]

@xiangjinwu @fuyufjh @yuhao-su I've made the changes, please take a review when you have time 😃