Update Document Metadata and labels after importing and vectorizing

Description

I imported a bunch of pdfs before understanding all the features of the tool. Now I’d like to use the metadata and labels to the documents in weaviate.
Is it possible to add metadata and labels after the documents are vectorized?
If so, how?

Server Setup Information

  • Weaviate Server Version:
  • Deployment Method:
  • Multi Node? Number of Running Nodes:
  • Client Language and Version:
  • Multitenancy?:

Any additional Information

hi @bam !!

It is possible. Note that, if you have not set that property to be skipped from vectorization, whenever you update an object, it will trigger a new vectorization of that object.

If you don’t want that, you can set the property you want to use as metadata/filtering needs to be skipped or not listed as the source_properties if using a named vector.

Let me know if this helps!

Thanks!

Thank you for the quick reply. Can this be done inside Verba? I followed the link. The instructions imply it may only be done by coding it with the object id and collection name. I’m not able to see the object id nor collection name in Verba.

hi! Verba do not provide a way to edit chunks, so I believe you will need to drop the entire document and reupload, or connect direct to Weaviate and change the objects you want.

AFAIK, Verba do not expose the ID or collection at UI level. :grimacing:

So you’ll need to list the collections, and look for the object you want to change.

Let me know if this helps!

1 Like