How do we query on a secondary index of dynamodb using boto3?

You need to provide an IndexName parameter for the query function.

This is the name of the index, which is usually different from the name of the index attribute (the name of the index has an -index suffix by default, although you can change it during table creation). For example, if your index attribute is called video_id, your index name is probably video_id-index.

import boto3
from boto3.dynamodb.conditions import Key
dynamodb = boto3.resource('dynamodb')
table = dynamodb.Table('videos')
video_id = 25
response = table.query(
    IndexName='video_id-index',
    KeyConditionExpression=Key('video_id').eq(video_id)
)

To check the index name, go to the Indexes tab of the table on the web interface of AWS. You'll need a value from the Name column.


Adding the updated technique:

    import boto3
    from boto3.dynamodb.conditions import Key, Attr

    dynamodb = boto3.resource(
         'dynamodb',
         region_name='localhost',
         endpoint_url='http://localhost:8000'
    )

    table = dynamodb.Table('userTable')

    attributes = table.query(
        IndexName='UserName',
        KeyConditionExpression=Key('username').eq('jdoe')
    )
    if 'Items' in attributes and len(attributes['Items']) == 1:
        attributes = attributes['Items'][0]


For anyone using the boto3 client, below example should work:

import boto3    

# for production
client = boto3.client('dynamodb')

# for local development if running local dynamodb server
client = boto3.client(
   'dynamodb',
   region_name='localhost',
   endpoint_url='http://localhost:8000'
)

resp = client.query(
   TableName='UsersTabe',
   IndexName='MySecondaryIndexName',
   ExpressionAttributeValues={
       ':v1': {
           'S': '[email protected]',
       },
   },
   KeyConditionExpression='emailField = :v1',
)

# will always return list
items = resp.get('Items')

first_item = items[0]

There are so many questions like this because calling dynamo through boto3 is not intuitive. I use dynamof library to make things like this a lot more common sense. Using dynamof the call looks like this.

from dynamof.operations import query
from dynamof.conditions import attr

query(
    table_name='users',
    conditions=attr('role').equals('admin'),
    index_name='role_lookup_index')

https://github.com/rayepps/dynamof

disclaimer: I wrote dynamof