I found that in general, it was OK for pretty obvious things but sometimes it missed stuff such as: running a LIKE with wild card at the start, ie: "LIKE '%Luke%'". Running that wildcard in the start tends to make the optimizer do a full scan.
Otherwise, it did okay. Just ensure you give ChatGPT the schema as well for its best suggestion.
I found that in general, it was OK for pretty obvious things but sometimes it missed stuff such as: running a LIKE with wild card at the start, ie: "LIKE '%Luke%'". Running that wildcard in the start tends to make the optimizer do a full scan.
Otherwise, it did okay. Just ensure you give ChatGPT the schema as well for its best suggestion.