I wrote a post a while ago looking at the performance of the T-SQL endpoint compared to using FetchXML. Now I’ve got a better understanding of how it works, I wanted to check if the same performance considerations I’ve previously encountered when working with the filtered views on-premise still apply.