Support of DynamoDB for Python

CAST supports DynamoDB via its com.castsoftware.python extension. Details about how this support is provided is discussed below.

Objects

Boto3 framework

The support for DynamoDB in boto3 is focused on creating DynamoDB tables and adding links representing CRUD-like operations on them.

Methods for creation of tables

Methods for getting a paginator from a DynamoDB client object

High-level DynamoDB resource methods

High-level table resource methods for the DynamoDB database API

Batch-writer methods

Paginator methods

Low-level client methods for the DynamoDB database API

What results can you expect?

Creation of DynamoDB tables

The two similar create_table() methods provided by boto3 are supported: for the low-level client object and the resource object (as in the example below):

# https://docs.aws.amazon.com/amazondynamodb/latest/developerguide/GettingStarted.Python.01.html

import boto3

def create_movie_table(dynamodb=None):
    if not dynamodb:
        dynamodb = boto3.resource('dynamodb', endpoint_url="http://localhost:8000")

    table = dynamodb.create_table(
        TableName='Movies',
        KeySchema=[...],
        AttributeDefinitions=[...],
    )
    return table

Upon resolution of the table name, a Python DynamoDB Table object is created under a Python DynamoDB Database:

Python DynamoDB Database objects are only created if dynamodb tables are present in the results. A single Python DynamoDB Database object is created per project at most. No link is created between the artifact invoking table creation and the resulting table object. However, the bookmark of the table does reflect the corresponding method call.

When the name of a table is not resolved (either because of absence of information or technical limitations) a Python Unknown DynamoDB Table is created instead. A single unknown table is created per project.

Actions on DynamoDB tables

Only actions that have an impact on data in DynamoDB tables are modelized. These actions are represented with CRUD links. In the example below the method batch_write_item() of the low-level client object (DynamoDB) is called. This method can insert new table items or delete them, therefore the presence of the two link types (insertUseLink, deleteUseLink).

# sample_batch_write_item.py
import boto3

TABLE = 'fabric'
dynamodb = boto3.client('dynamodb')

def write_item():
    response = dynamodb.batch_write_item(
        RequestItems = {
          TABLE: [{"PutRequest": { "Item": { ... } }}]
        }
    )

Known Limitations

• Accessing tables via tables collections such as “dynamodb.tables.all()” is not supported.
• The query language PartiQL is not supported (and thus related API calls listed below).
• The following methods might be supported in the future: