DynamoDB

Introduction

DynamoDB is Aamazon's distributed key-value store, optimized for fast read and writes of data. Like many other distributed key-value stores, its query language does not support joins.

It can run in either Strongly Consistent mode (similar to HBase), or Eventually Consistent mode (like Cassandra).

Preliminaries

When using AWS managed services, the boto library for Python is a good tool to connect and setup AWS services.

To install boto, run the command

pip install boto3

In order to not have your AWS credentials in your source code, boto allows you to store them in a file ~/.boto, which you have to create.

## copy and paste the three lines (modified with your AWS keys) into ~/.boto
[Credentials]
aws_access_key_id=XXXXX
aws_secret_access_key=XXXxxxXXX/XXXxxxXX

Example creating a table.

Using boto, a simple script to create a table in DynamoDB is the following

import boto3

dynamodb = boto3.resource('dynamodb')

table = dynamodb.create_table(
    TableName='myTable',
    KeySchema=[
        {
            'AttributeName': 'myKey',
            'KeyType': 'HASH'
        }
    ],
    AttributeDefinitions=[
        {
            'AttributeName': 'myKey',
            'AttributeType': 'S'
        }

    ],
    # pricing determined by ProvisionedThroughput
    ProvisionedThroughput={
        'ReadCapacityUnits': 5,
        'WriteCapacityUnits': 5
    }
)

The key features of this script are that it contains a key, 'myKey' this example, which is specified as a HASH key and is a String (hence the S type). Note that it is not needed to specify a schema for each row, which is why DynamoDB is sometimes referred to as a "schemaless database".

Note that the last configs, 'ReadCapacityUnits' and 'WriteCapacityUnits' determine how many reads/writes are per second.

Example to write data into DynamoDB

An example script to write data to the table we created is

import boto3

dynamodb = boto3.resource('dynamodb')

table = dynamodb.Table('myTable')

for i in range(0,100):
        curKey = "key" + str(i)
        curValue = "Hello World! This is data entry  " + str(i)
        response = table.update_item(
                Key={
                        "myKey": curKey
                },
                UpdateExpression="set myValueString = :val",
                ExpressionAttributeValues={
                        ':val':curValue
                },
                ReturnValues="UPDATED_NEW"
        )