So don't use an auto_increment field. Use a date_time field instead (call it something like "create_date"), and whenever a new row is inserted, set its value with the "NOW()" function. Then you'll be able to "order by create_date desc limit 10".
You don't need an auto_increment field -- you just need "url" to be the primary key, and you need something chronological to sort on.
Update: About the 2nd part of your initial question: you could try to work out a two-stage query: find the "create_date" value for the row that is currently 10 rows "back" from the most recent row, then select all rows where create_date is greater than that value, and order by "Total".
But I think it would be a lot easier to do a single query ordered chronologically, store those into an AoA or AoH and let the perl script sort them based on "Total".
In reply to Re^3: UPDATE ON DUPLICATE and sorting
by graff
in thread UPDATE ON DUPLICATE and sorting
by coldfingertips
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |