Skip to content

Latest commit

 

History

History
executable file
·
184 lines (112 loc) · 4.83 KB

README.rst

File metadata and controls

executable file
·
184 lines (112 loc) · 4.83 KB
Raw

Amazon DAX Client for Python

The Amazon DAX Client for Python is used to access Amazon DAX clusters from Python. It is nearly source-compatible with Boto3, with only a small change needed to the client initialization to use DAX instead of DynamoDB. Creating a connection to your DAX cluster requires using the cluster discovery endpoint URL returned in the DescribeClusters response as the endpoint.

Installation

Install Amazon DAX Client using pip:

$ pip install amazon-dax-client

Quick Start

Boto3 has two different interfaces, the resource interface and the botocore client interface. Both are supported by the Amazon DAX client, with slightly different client initialization.

For the resource API, change from:

ddb = boto3.resource('dynamodb')

to

dax = AmazonDaxClient.resource(
    endpoint_url=<cluster_discovery_endpoint_url>)

All other boto3.resource() arguments are accepted.

For the botocore client API, change from:

session = botocore.session.get_session()
ddb = session.create_client('dynamodb', ...)

to

session = botocore.session.get_session()
dax = AmazonDaxClient(session, ...)

For Boto3 client API, change from:

ddb = boto3.client('dynamodb')

to

dax = AmazonDaxClient(endpoint_url=<cluster_discovery_endpoint_url>)

The Boto3 and botocore client APIs are exactly the same.

Hostname Verification for TLS connections is enabled by default when making requests from the client to the cluster and has no effect for unencrypted clusters. This can be turned off using the client API, be sure you understand the implication of turning it off, which is the inability to authenticate the cluster that you are connecting to.

Unsupported Features

The Amazon DAX client does not support table operations. Any table manipulation operations must use the regular Boto3 or botocore DynamoDB client.

Paginators are not currently supported for DAX.

Documentation

Once created, the interface is the same as the Boto3/botocore DynamoDB clients.

For acomplete example, follow the guide to create a sample app.

Getting Help

Please use these community resources for getting help.

Changes

v2.0.0

  • Encryption in transit feature
  • Drop support for python versions below 3.6

v1.1.8

  • Retry logic bug fixes

v1.1.7

  • Minor bug fixes

v1.1.6

  • Linting fixes
  • Updated logging to log on Logger objects

v1.1.5

  • Add jitter and backoff for retryable exceptions

v1.1.4

  • Fixes bug starting with clusters of less than 3 nodes.

v1.1.3

  • Fixes and improvements to service discovery logic.
  • Add support for Python 3.8.
  • Fix 'RetryHandler' object has no attribute '_cluster': AttributeError problem reported in the AWS Forums.
  • Fix problems found by static analysis.

v1.1.2

  • Fixes a bug that can result in a failure to update the roster when it changes, which can later result in a NoRouteException.

v1.1.1

  • Fixes a bug that can result in an infinite loop on node failure.
  • Update TransactWriteItem test item limit to 25.

v1.1.0

  • Adds support for transact-get-items and transact-write-items APIs for DyanamoDB transactions.
  • Improved efficiency of connection pooling.

v1.0.7

  • Fix scheduling of background tasks.

v1.0.6

  • Properly de-anonymize UnprocessedItems results from BatchWrite.
  • Raise a proper error if no backends are available.

v1.0.5

  • Fix UpdateItem result parsing with a subset of attributes changed

v1.0.4

  • Use user-specified timeouts
  • Fix issue with update response if the item is not changed
  • Fix error decoding BatchWrite UnprocessedItems.

v1.0.3

  • Fix AmazonDaxClient.resource() when using batch_get_items or batch_write_items.

v1.0.2

  • Fix Python 2 encoding issues
  • Fix decoding of ConsumedCapacity, ItemCollectionMetrics in batch operations

v1.0.1

  • Initial release