Question

so suppose I have this query:

SELECT * FROM table1 A, table1 B where A.id = B.id;

even if there is an index on id on table1, it would still scan the entire table…is there a way to speed this up so that it doesn’t scan the entire table?

Answer 1

Though I do not have your execution plan in front of my, this is actually only scanning TableA most likely, and the reason for that is because you have no other conditions on TableA. It should be seeking on TableB at this time just because it’s statistical. Now, if you were to provide another condition for TableA, and that condition met an index, you would find that it was not scanning the entire table.

To further that, if the index on TableA was a covered index for the query it would actually never even read a data page for TableA.

And to further that, if there were a foreign key constraint between TableA and TableB and you weren’t reading anything more from TableB than the id field it would also never read a data page.

And to further that, even if you did need data from TableB, if that data were on an index it still wouldn’t have to read a data page.

But finally, because you’re selecting *, it’s unlikely this will ever be a very efficient query because it’s selecting all columns from both tables.

I hope that wasn’t too long winded but I wanted to make sure that you knew where I was coming from.