This article will discuss the advantages and disadvantages of using namespaces and metadata filtering in your application. Performance is the same whether you use namespaces or metadata filtering. The most significant difference is how you query your index.
Namespaces
Namespaces are a way to segment data into distinct areas within your index. The intent is to have the ability for an index to serve multiple purposes. For example, you could have a single index containing customers, catalog items, and articles. These can be queried and treated effectively as separate entities in your index. However, there is one strong consideration. You can only query one namespace (or none) at a time. This means you cannot choose to query the entire corpus of data in the future. If you see the need to query across namespaces, then use metadata filtering instead. If you never need to cross namespaces with queries, then using namespaces is fine.
See also namespaces
Metadata Filtering
Metadata fields or you could call them key:value pairs, are a way to add information to individual vectors to give them more meaning. By adding metadata to your vectors, you can filter by those fields at query time. This is similar to namespaces, except you are not limited to a single filter. You can use a variety of filter patterns and conditions to search subsets of your data without requiring namespaces or multiple queries. This is a popular alternative to namespaces, and many customers use this method instead. This gives the same performance and more flexibility in the future if you want to search across the entire index.
See also metadata filtering
Switching from namespaces to metadata filtering:
1.) Clear your index or create a new index
2.) Re-ingest your data and use a metadata field instead. Leave out the namespace parameter.
3.) Now you can run queries using these methods: https://www.pinecone.io/docs/metadata-filtering/
Comments
0 comments
Please sign in to leave a comment.