This code change will change the list of identical "Column type TIMESTAMP_LTZ is not supported for conversion from snowflake to looker. No dimension will be created." to start with the column names that aren't being converted.

This is useful to me and my peers because it tells us which columns need to be manually written and described in the LookML file. It continues to allow map_adapter_type_to_looker() to work as it currently does in case there is no column name (not sure this matters though!).

I've tested this by running the current prod dbt2looker command against a file in my company's codebase, then running the modified dbt2looker. More testing in the reviewer's environment wouldn't hurt.

This would be even better if model name were passed to the map_adapter_type_to_looker() function too to allow for the even clearer column name pattern of model_name.column_name