Skip to main content

Dynamodb

The Dynamodb source allows you to sync data from Dynamodb. The source supports Full Refresh and Incremental sync strategies.

Resulting schema

Dynamodb doesn't have table schemas. The discover phase has three steps:

Step 1. Retrieve items

The connector scans the table with a scan limit of 1k and if the data set size is > 1MB it will initiate another scan with the same limit until it has >= 1k items.

Step 2. Combining attributes

After retrieving the items it will combine all the different top level attributes found in the retrieved items. The implementation assumes that the same attribute present in different items has the same type and possibly nested attributes values.

Step 3. Determine property types

For each item attribute found the connector determines its type by calling AttributeValue.type(), depending on the received type it will map the attribute to one of the supported Airbyte types in the schema.

Features

FeatureSupported
Full Refresh SyncYes
Incremental - Append SyncYes
Replicate Incremental DeletesNo
NamespacesNo

Full Refresh sync

Works as usual full refresh sync.

Incremental sync

Cursor field can't be nested, and it needs to be top level attribute in the item.

Cursor should never be blank. and it needs to be either a string or integer type - the incremental sync results might be unpredictable and will totally rely on Dynamodb comparison algorithm.

Only ISO 8601 and epoch cursor types are supported. Cursor type is determined based on the property type present in the previously generated schema:

  • ISO 8601 - if cursor type is string
  • epoch - if cursor type is integer

Getting started

This guide describes in details how you can configure the connector to connect with Dynamodb.

Role Based Access

Defining access_key_id and secret_access_key will use User based Access. Role based access can be achieved by omitting both values from the configuration. The connector will then use DefaultCredentialsProvider which will use the underlying role executing the container workload in AWS.

Сonfiguration Parameters

  • endpoint: aws endpoint of the dynamodb instance
  • region: the region code of the dynamodb instance
  • (Optional) access_key_id: the access key for the IAM user with the required permissions. Omit for role based access.
  • (Optional) secret_access_key: the secret key for the IAM user with the required permissions. Omit for role based access.
  • reserved_attribute_names: comma separated list of attribute names present in the replication tables which contain reserved words or special characters. https://docs.aws.amazon.com/amazondynamodb/latest/developerguide/Expressions.ExpressionAttributeNames.html

Changelog

Expand to review
VersionDatePull RequestSubject
0.3.32024-07-22PR_NUMBER_PLACEHOLDERfix primary key fetching
0.3.22024-05-0127045Fix missing scan permissions
0.3.12024-05-0131935Fix list more than 100 tables
0.3.02024-04-2437530Allow role based access
0.2.32024-02-1335232Adopt CDK 0.20.4
0.2.22024-01-2434453bump CDK version
0.2.12024-01-03#33924Add new ap-southeast-3 AWS region
0.2.018-12-2023https://github.com/airbytehq/airbyte/pull/33485Remove LEGACY state
0.1.201-19-2023https://github.com/airbytehq/airbyte/pull/20172Fix reserved words in projection expression & make them configurable
0.1.102-09-2023https://github.com/airbytehq/airbyte/pull/22682Fix build
0.1.011-14-2022https://github.com/airbytehq/airbyte/pull/18750Initial version