A tool to duplicate DynamoDB tables.

duplicate_dynamodb_table.py

from traceback import print_exc
from boto.dynamodb2.exceptions import ValidationException
from boto.dynamodb2.fields import HashKey, RangeKey
from boto.dynamodb2.layer1 import DynamoDBConnection
from boto.dynamodb2.table import Table
from boto.exception import JSONResponseError
from time import sleep
import sys

if len(sys.argv) != 3:
    print 'Usage: %s <source_table_name> <destination_table_name>' % sys.argv[0]
    sys.exit(1)

src_table = sys.argv[1]
dst_table = sys.argv[2]
DynamoDBConnection.DefaultRegionName = "us-west-2"
ddbc = DynamoDBConnection(aws_access_key_id='AKlalala',
                          aws_secret_access_key='SDFSDFSDFSDF')


# 1. Read and copy the target table to be copied
table_struct = None
try:
    sourceTable = Table(src_table, connection=ddbc)
    table_struct = sourceTable.describe()
except JSONResponseError:
    print "%s not existing" % src_table
    sys.exit(1)

print '*** Reading key schema from %s table' % src_table
src = ddbc.describe_table(src_table)['Table']
hash_key = ''
hash_key_type = ''
range_key = ''
range_key_type = ''
for schema in src['KeySchema']:
    if schema['KeyType'] == 'HASH':
        hash_key = schema['AttributeName']
        hash_key_type = (some_dict['AttributeType'] for some_dict in src['AttributeDefinitions'] if some_dict['AttributeName'] == hash_key).next()
    elif schema['KeyType'] == 'RANGE':
        range_key = schema['AttributeName']
        range_key_type = (some_dict['AttributeType'] for some_dict in src['AttributeDefinitions'] if some_dict['AttributeName'] == range_key).next()
schema = [HashKey(hash_key, data_type=hash_key_type)]
if range_key:
    schema.append(RangeKey(range_key, data_type=range_key_type))

# 2. Create the new table
table_struct = None
try:
    targetTable = Table(dst_table, schema=schema, connection=ddbc)
    table_struct = targetTable.describe()
    print 'Table %s already exists' % dst_table
    sys.exit(0)
except JSONResponseError:
    targetTable = Table.create(dst_table, schema=schema, connection=ddbc)
    print '*** Waiting for the new table %s becomes active' % dst_table
    sleep(2)
    while ddbc.describe_table(dst_table)['Table']['TableStatus'] != 'ACTIVE':
        sleep(2)

# 3. Add the items 
for index, item in enumerate(sourceTable.scan()):
    print "Copying item #%d..." % (index + 1)
    new_item = {}
    new_item[hash_key] = item[hash_key]
    if range_key != '':
        new_item[range_key] = item[range_key]
    for f in item.keys():
        if f in [hash_key, range_key]:
            continue
        new_item[f] = item[f]
    try:
        targetTable.put_item(new_item, overwrite=True)
    except ValidationException:
        print_exc()
        print dst_table, new_item
    except JSONResponseError:
        print ddbc.describe_table(dst_table)['Table']['TableStatus']

Fixed lots of concrete table specific code in the original, added support for different range/hash key types, and tables with no range key.