Kafka as a database or even as a cache

I am wondering whether anyone has used Kafka as a database? To maintain a database like behavior with insert/read/update/delete operations. I know compacted topics could be used to behave like an update operation. With KTable and GlobalKTable seems to give more control on accessing the data. How has anyone used it if its been used partly like a database with some performance metrics on how fast or slow it is? Also, has anyone tried to use Kafka as a cache with materialized views for each service syncing with changes as they come in? What about performance and feasibility? Please share your thoughts and experience if any.