Copy Namespace Feature

[Ahtesham00]

Is there any copy namespace feature available in pinecone?

I have tried to search about it but have not been able to find any sort of solution provided by pinecone SDK.

It seems like the only way to copy the namespace is to get all the results and copy it and re upload it.

[igorbilyk2]

Also interested in that feature

[anawishnoff]

Hey @igorbilyk2 and @Ahtesham00 , thanks for the feedback. Can you expand more on:

• What motivated you to submit this feature request/what problem this would solve?
• A clear and concise description of what you want to happen.
• Any alternatives you've considered?

[Ahtesham00]

@anawishnoff

1. What motivated you to submit this feature request/what problem this would solve?

My software users want the ability to copy the namespace or the data I am storing under a specific namespace so they can have it replicated elsewhere in the system. The issue isn't with copying itself, as I can point my users to the same vectors. The problem arises when they want to expand on the copied namespace. For example, if I have 100 vectors in namespace X, and my application users want to copy this namespace, they should be able to upload more vectors to the copied namespace.

2. A clear and concise description of what you want to happen.

I would like to have a feature that allows users to duplicate an existing namespace along with its vectors and then enable them to add more vectors to the duplicated namespace. This functionality should allow the copied namespace to be independent of the original, enabling users to expand on it as needed.

3. Any alternatives you've considered?

In the past, I've used an alternative method where I extract all the results and then reupload them. However, this approach has its limitations. Specifically, I have to assign a specific top_k value during the extraction, which can change over time and may not be optimal for all scenarios.

[anawishnoff]

Hey @Ahtesham00 , thanks for providing those details! Have you looked into collections (for pod-based indexes) or backups (launched recently for Serverless indexes)? You can find the documentation here: https://docs.pinecone.io/guides/indexes/understanding-backups-and-collections. This allows you to create copies of your indexes, and reupload it somewhere where it can be added to/modified.