I would like to use the $rowid$ from a column table as foreign key to join that table with another table. I thought that this would performing great is its a simple numeric id. When I looked at the query plan of the a statement which used $rowid$ it showed the join condition as a non-field join condition and the performance was really bad. What is happening here and is there a way to optimize the usage of the row pointer. Another question related is if the the row pointer is persistent in a table or if it drifts over time. I performed some inserts and deletes after the row pointer was unchanged.
↧