Database / BetterDB Interview questions
How does the Search Query Runner build hybrid KNN queries?
The panel offers a visual filter builder for TAG, NUMERIC, and TEXT fields – toggles, ranges, and multi-select tag pickers – alongside a separate vector KNN search section with a K selector and a vector input, entered either as a JSON array like [0.12, -0.4, ...] or as base64-encoded FLOAT32 bytes.
When you use both together, BetterDB automatically combines the visual filter with a =>[KNN K @field $vec] clause to produce a single hybrid FT.SEARCH command, and the command preview updates live so you can copy it straight into valkey-cli or a script.
The panel is also capability-aware: it checks whether the connected server is running valkey-search or RediSearch and hides controls, like GEOSHAPES filtering, that the detected engine doesn't actually support, instead of letting you build a query that would just error when run.
More Related questions...
