Which implementation works faster, ADO or ODBC? Is there any relevant information (or benchmarks) comparing the difference between linking directly to the ODBC or accessing databases through the ADO components?
I haven’t timed them so I can’t say. There’s also another variable that could weigh in your choice.
My understanding is that ADO is built on OLE DB and so has the capacity to manipulate nontabular data, in contrast to ODBC, which is only for database tables.
I wouldn’t approach the choice from the dimension of speed. It has more to do with flexibility in handling nondatabase data (like project files or documents), not to mention that ADO’s object model gives you much more control over the object and easier trappable error information about your process.