The post asks about the new release of the Pinecone index, which allows a single Pinecone index to be shared among multiple vector indices. A community member explains that to do this, you need to define a metadata_filter variable when initializing the index. Another community member asks about the difference between this solution and using a composed graph with a keyword table index, and the first community member clarifies that with the metadata_filter approach, you can use the same underlying Pinecone index structure across multiple vector indices, whereas a composed graph with a keyword table index would be a different approach.
Hey! I saw the new release about the expansion of pinecone index. It says that a single pinecone index is sharable among multiple vector indices. Anyone able to explain more?
@AndreaSel93 if you specify metadata_filter, this just means that you can use the same underlying pinecone.Index structure across multiple vector indices. the behavior of the query call is unchanged (the same as if you queried any vector index). that's diff than defining a composed graph with a keyword table index
