Dirk Dunger wrote:
> Hi!
> I use the program testReadPerf to do some benchmarking on NDB Cluster
> and when I tried to make sense of the timings I get, I realized that I
> could use some insight on what that thing is doing in it's individual
> tests.
>
> I'm wondering what exactly the different tests in this benchmark are.
> From the MySQL reference manual I know what serial/batch pk and
> serial/batch uniq index do, but I'm curious what "eq-bound" means and
> what kind of index is used in the "index eq-bound" and "index range" tests.
> In essence the tests "index eq-bound", "index range" and "index ordered"
> are some kind of mystery to me.
>
> It would be really nice if someone here could explain these things.
> (maybe with some mysql query like example, if that is possible)
>
> Thanks
> Dirk
>
Hi,
create table t1 (a int primary key, b int, c int, unique key using hash(b), key (c))
engine = ndb;
serial pk: select * from t1 where a = ?
batch pk: select * from t1 where a in (?, ?, ?, ...)
serial unique: select * from t1 where b = ?
batch unique: select * from t1 where b in (?, ?, ?, ...)
eq-bound: select * from t1 where b = ?
index range: select * from t1 where b between ? and ?
/jonas
ps.
not sure that syntax is correct
ds.
