This is almost as depressing as the heat-death of the universe.
jazzbass
Databases usually support multiple isolation levels: serializable, repeatable reads, read committed and read uncommitted. With the highest level of isolation, there is a big loss in concurrency, but with a low level of isolation you encounter problems such as phantom reads, nonrepeatable reads and dirty reads.
I wonder, under which isolation level where these tests performed?
This is almost as depressing as the heat-death of the universe.
Databases usually support multiple isolation levels: serializable, repeatable reads, read committed and read uncommitted. With the highest level of isolation, there is a big loss in concurrency, but with a low level of isolation you encounter problems such as phantom reads, nonrepeatable reads and dirty reads.
I wonder, under which isolation level where these tests performed?
More information about isolation levels here.