Find answers from the community

Updated last year

i need help..

i need help..
whats the difference between index store, vector store, document store and graph store? is there like a deep dive i can watch/read to understand better?
e.g. i've been playing with milvus and pgvector. so the vector store component is kinda clear.
but what is an index store, document store and graph store? if i'm building a production grade app, do i need the index store, document store and graph store ?
Thank you in advance!

4 comments

WWhiteFang_Jr

Here this might help you understanding these terms: https://docs.llamaindex.ai/en/stable/module_guides/storing/storing.html

aaristokratic.eth

in https://docs.llamaindex.ai/en/stable/module_guides/storing/vector_stores.html
it states that postgres supports metadata filtering, hybrid search, delete, store documents and async.
i assume when it says store documents, it means that for the docstore, we can point it to postgres?
cant seem to find any examples of these..

Attachment

WWhiteFang_Jr

For all the third party Vector stores, I think docs and everything are stored along with with vectors, Sharing the code example for postgres:
https://github.com/run-llama/llama_index/blob/773e7dfa484a15b3c8e5e4758bacc32f1be0f591/llama_index/vector_stores/postgres.py#L283

aaristokratic.eth

Thanks @WhiteFang_Jr -

Add a reply