The performance and partly the output of the code is always dependent on the table definition. * instead of column names just removes an output limiter, which can be useful or can be irrelevant, depending on the context.

Though sure, known to negatively affect performance, I think in some database systems more than in others?