Retrieving vectors

Vector Databases for Embeddings with Pinecone

James Chapman

Curriculum Manager, DataCamp

Recap...

An index being created.

Recap...

Several records being ingested into the index. The index fullness increases once the records are added.

Accessing vectors

Fetching

Retrieve vectors based on their IDs

A book being retrieved based on its ISBN.

Querying

Retrieve similar vectors to an input vector

Amazon book recommendations.

Fetching vectors

index.fetch(

    ids=['0', '1']

)

{'namespace': '',
 'usage': {'read_units': 1},
 'vectors': {'0': {'id': '0',
                   'metadata': {"genre": "productivity", "year": 2020},
                   'values': [0.025525547564029694, ...]},
             '1': {'id': '1',
                   'metadata': {"genre": "action", "year": 2023},
                   'values': [-0.0131468913, ...]}}
}

¹ https://docs.pinecone.io/guides/data/fetch-data

Read units

Measure of the resources consumed during read operations:
- Fetching → 1RU / 10 records
- Querying
- Listing

Pinecone starter account pricing.

¹ https://www.pinecone.io/pricing/

Fetching vectors from namespaces

index.fetch(
    ids=['0', '1']

    namespace='namespace1'

)

{'namespace': 'namespace1',
 'usage': {'read_units': 1},
 'vectors': {'0': {'id': '0',
                   'metadata': {"genre": "productivity", "year": 2020},
                   'values': [0.025525547564029694, ...]},
             '1': {'id': '1',
                   'metadata': {"genre": "action", "year": 2023},
                   'values': [-0.0131468913, ...]}}
}

Let's practice!

Vector Databases for Embeddings with Pinecone