[Feature Request]: insert faster, general documents are only saved to vectorDB, only important documents are saved to vectorDB+graphDB

[zhouzhou12]

Do you need to file a feature request?

• I have searched the existing feature request and this feature request is not already filed.
• I believe this is a legitimate feature request, not just a question or bug.

Feature Request Description

For large scale data, in order to insert faster, general documents are only saved to vectorDB, only important documents are saved to both vectorDB and graphDB.

Additional Context

No response

[danielaskdd]

The default hybrid query mode currently only searches the vector database, which may result in incomplete document retrieval if only important documents are saved to both vector and graph DB . Could you share more about your specific use case requirements?

[zhouzhou12]

We have a large knowledge data, maybe more than 1GB. We use a 14B-parameter LLM on two GPUs.
If saving all the data to both vector database and graph database, it will take a very long time.
If the data is divided into important and unimportant data, with the unimportant data stored only in the vector database, and the important data stored in both vector database and graph database, a significant amount of time will be saved.
During queries, a "mix" mode can be used: important data is retrieved from both vector database and graph database, while unimportant data is retrieved only from the vector database. This approach should already be sufficient in terms of effectiveness.

The default hybrid query mode currently only searches the vector database, which may result in incomplete document retrieval if only important documents are saved to both vector and graph DB . Could you share more about your specific use case requirements?