|
|
|
|
|
|
by hamilyon2
1182 days ago
|
|
|
> For example, if you are 100% sure that two transactions would never touch the same row, or that they’d operate on different columns - you can always consider easing the isolation I always thought of database row as an immutable tuple. It is not like row-oriented databases will allow you to operate on disjoint sets of columns as if they were different tables. Lower isolation levels lower than read committed will introduce same unexpected artifacts even if column sets don't intersect |
|
|