On Fri, Aug 4, 2017 at 3:24 AM, Fabien COELHO <coelho@cri.ensmp.fr> wrote:
>
>>> For the CREATE stuff, the script language is SQL, the command to use it
>>> is
>>> "psql"...
>>
>>
>>> The real and hard part is to fill tables with meaningful pseudo-random
>>> test data which do not violate constraints for any non trivial schema
>>> involving foreign keys and various unique constraints.
>>
>>
>>> The solution for this is SQL for trivial cases, think of:
>>> "INSERT INTO Foo() SELECT ... FROM generate_series(...);"
>>
>>
>> Yeah. I was also thinking that complicated data-generation requirements
>> could be handled with plpgsql DO blocks, avoiding the need for hard-wired
>> code inside pgbench.
>
>
> I do not thing that it is really be needed for what pgbench does, though.
> See attached attempt, including a no_foreign_keys option.
>
> The only tricky thing is to have the elapsed/remaining advancement report on
> stdout, maybe with some PL/pgSQL.
>
> Timings are very similar compared to "pgbench -i".
>
The generating data with plpgsql DO blocks means that we do the
data-generation on sever side rather than on client side. I think it's
preferable in a sense because could speed up initialization time by
reducing the network traffic.
Regards,
--
Masahiko Sawada
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center