Self Query is one of the most important use cases for production vector store retrieval, adding ability to filter on metadata (columns) and leverage predicate pushdown. At the moment cloud spanner doesn't support it and trying something like
https://python.langchain.com/v0.1/docs/modules/data_connection/retrievers/self_query/
yields a 'not supported' error. From my cursory look, however, it seems to be a relative easy lift, granted how PGVector and other similar RDBMS-like databases already do it in langchain.