Pretty self-explanatory. See the screenshot above. I'm using a newly deployed 3 node cluster, with 5.15.0 installed. I'm following the docs closely, so I don't think I've made a syntax mistake with this code
CREATE VECTOR INDEX section_content_e
FOR (n:Section) ON n.content_embedding_euclidean
OPTIONS {indexConfig:{
'vector.dimensions':1536,
'vector.similarity_function':'euclidean'
}}
If I use the deprecated method for vector index creation (ie. CALL db.index.vector.createNodeIndex()), then I can create indexes as expected.
Anyone have any ideas?