There is no need to apologize. You have been very good at describing both your original question, and giving good and relevant information along the way, which makes a world of difference.
Anyways, normally new questions go in new threads. Though I dont really know if thats a new question.
Now from a design point of view, I wouldn't choose the path you are choosing, unless you have very few different road names. Just imagine 20 road names, gives you 20 tables, and then making 20 queries, and if say the txt file gets another field, you have to update all 20 tables (without making human errors...) and the queries.
Now I would do:
1) Import file to temporary table.
2) Get road name. Check if its a known road name, or new road name. If new, insert into table of road names. Get the Primary Key of the tbl_RoadNames.
3) Append the temporary table+Primary Key into the main table, in which there is a Foreign Key field for the road ID.
That gives you all your data in one table (less maintenance) and makes it easier to query.
You can do step 2 as well for the dataset ID.
I hope that was clear enough.