Reputation: 51
I'm building a RAG chatbot using Langchain, using the data I've stored in a Qdrant vector database.
I wanted to change the metadata of a few documents in my qdrant vector database.
For this, I stored those documents separately (including the vectors), deleted the old data, modified the metadata and upserted them back, without embedding them again. I want to know whether it is safe to do that. (Note: I only changed the metadata. The page content is the same.)
Just for more context, when I scrolled the points, a record looked something like this:
Record(id='001c7e73032a40158b0c629f163e3bcf', payload={'page_content': 'The cat (Felis catus), commonly referred to as the domestic cat or house cat, is a small domesticated carnivorous mammal. It is the only domesticated species...', 'metadata': {'source': 'Wikipedia', 'url': 'https://en.wikipedia.org/wiki/Cat', 'count': 1}}, vector=[-0.013379335966146134, 0.007072219895927447, -0.0018562082621154511, -0.0036226000104401957, -0.005558645199573012, 0.012487823730867539, -0.023378909852710315,...
So my question is: I have retained the vectors, page_content and ids. Does this eliminate the need for re-embedding the data?
PS: The reason behind avoiding the re-embedding of data is the high cost associated with embedding the data, since i have a lot of data.
Upvotes: 1
Views: 222