Caching AppendData plan causes data to be inserted twice

We’ve identified an issue where a DataFrame created from an INSERT SQL statement and then cached will cause the INSERT to be executed twice. This happens because the logical plan for the INSERT (AppendData) doesn’t extend the IgnoreCachedData trait, so it isn’t ignored during caching as expected. As a result, the plan is cached and re-executed. We should fix this by ensuring that plans used by INSERT all extend the IgnoreCachedData trait.