T-SQL: delete one row without beeing able to identify it
It happened in a development database some time ago: I had more than one row with exactly the same values in any column including the ID column. I noticed the ID column (int) had no identity specification and changing that failed because of the duplicates.
The solution was to use the Top(n) syntax. You cannot only do
1 | select top(1) * from mytable where id=59 |
but also
1 | delete top(1) from mytable where id=59 |
That worked suprisingly well and I could setup the identity column after this