BynamoDB

High-Level DynamoDB Interface for Python wrapping Low-Level Interface of boto

Installation

From PyPi:

$ pip install bynamodb

From GitHub:

$ pip install git+https://github.com/teddychoi/BynamoDB#egg=bynamodb

Patch DynamoDBConnection

Set DynamoDBConnection default host and port.

from bynamodb.patcher import patch_dynamodb_connection

patch_dynamodb_connection(host='localhost', port=8000)

Model Definition

import datetime
from bynamodb.attributes import StringAttribute, StringSetAttribute
from bynamodb.indexes import GlobalAllIndex
from bynamodb.model import Model

class Article(Model):
    published_at = StringAttribute(hash_key=True)
    id = StringAttribute(range_key=True)
    title = StringAttribute()
    content = StringAttribute()
    author = StringAttribute()
    write_time = StringAttribute(
        default=lambda: str(datetime.datetime.now()))
    tags = StringSetAttribute(default=set())
    thumbnail = StringAttribute(null=True)

    class AuthorIndex(GlobalAllIndex):
        read_throughput = 5
        write_throughput = 5
        hash_key = 'author'
        range_key = 'published_at'

Put Item & Get Item

Article.put_item(
    published_at='2014-12-09',
    id='1',
    title='This is the title',
    content='This is the content',
    author='Bochul Choi'
)
article = Article.get_item(hash_key='2014-12-09', range_key='1')

Get Item from Raw Data

You can get items from raw data retrieved from boto's low level API

from boto.dynamodb2.layer1 import DynamoDBConnection

conn = DynamoDBConnection()
raw_data = conn.get_item(
    'Article',
    {
        'published_at': {'S': '2014-12-09'},
        'id': {'S': '1'}
    }
)
article = Article.from_raw_data(raw_data['Item'])

Simple Scan & Query

# Scan all articles that the title starts with "Title"
articles = Article.scan(title__startswith='Title')
# Get the total count of matching items
count = articles.count()
# Get the iterator of matching items
items = iter(articles)

# Query articles that author is "Bochul Choi"
articles = Article.query(author__eq='Bochul Choi', index_name='AuthorIndex')
# The query result also provide the count method and iterator
count = articles.count()
items = iter(articles)

Complex lookups in Scan & Query

from bynamodb.filterexps import Contains, GT

keyword = 'bynamodb'
filter_exp = GT('published_at', '2014-12-01') & (
    Contains('title', keyword) | Contains('content', keyword.upper()))

# Scan all articles that match the filter expression
articles = Article.scan(filter_exp)

# Query articles that match the filter expression and the author condition
author = 'Bochul Choi'
articles = Atricle.query(author__eq=author, filter_builder=filter_exp,
                         index_name='AuthorIndex')

Batch Writing & Batch Reading

with Article.batch_write() as batch:
    batch.put_item({
        'published_at': '2015-02-23',
        'id': '1',
        'title': 'Article 1',
        'content': 'This is the content',
        'author': 'Bochul Choi'
    })
    batch.put_item({
        'published_at': '2015-02-23',
        'id': '2',
        'title': 'Article 2',
        'content': 'This is the content',
        'author': 'Bochul Choi'
    })

articles = Article.batch_get(
    ('2015-02-23', '1'),
    ('2015-02-23', '2'),
)

Changelog

0.1.7

• Skip decoding non-existent attribute.

0.1.6

• Fix ResultSet to stop fetching items regardless of the limit param.

0.1.5

• Add support scan_index_forward and limit pamareter for Query API.

0.1.4

• Set ExclusiveStartKey as the untransformed LastEvaluatedKey.

0.1.3

• Model.delete needs to use encoded key.

0.1.2

• Model.put_item returns created item.

0.1.1

• NumberAttribute is only saved as number in db.