Hello all, we're using Typesense for a client to i...
# community-help
s
Hello all, we're using Typesense for a client to implement search of a catalog of file & web (i.e., URL) Resources that combines two types of data: • Resource catalog: Title, Summary, Organization, Categories, Type (web or file), Content Type, etc. • Full-text content: The full-length text content of file Resources (but not web) will be broken down into fragments (with page numbers), for granular & relevant full-text search. Target search UX: We're aiming for faceted browsing & navigation across all Resources (similar to Ecommerce Storefront) combined with full-text searching of all text, including the fragments, where one or more relevant fragments (with page numbers) can also appear with each matching Resources. Preferred approach: We'd like a normalized model with separate Resources and Resource Fragments collections to avoid data duplication and optimize maintenance. Limitations we encountered: - Can't query joined fields (#1657) - Can't facet on joined reference fields (#1706) Current fallback: Denormalized "Resource Fragments" collection where each fragment duplicates all parent Resource metadata, then group results by Resource ID to return one result per Resource. Are there any other solutions or approaches that we're missing, or is denormalization + grouping the only approach for this use case? And if so, are there any performance or functional issues we should know about? Any guidance on this is greatly appreciated. 🙏 Thanks much.
hello Typesense team & community, following-up to see if anyone can assist with our question. would appreciate any insights or guidance you can provide. thanks again.