Re: Idea to enhance pgbench by more modes to generate data (multi-TXNs, UNNEST, COPY BINARY)

Hi all,

Please take a look at https://commitfest.postgresql.org/patch/7387/ which
splits pgbench into functional modules so it would be easier to extend.

My aim is to re-base https://commitfest.postgresql.org/patch/7225/ on top
of that, and then add more useful things

Looks like I did not send the latest kleanup after Zsolt Parragi
<zsolt(dot)parragi(at)percona(dot)com> has some comments, will do it soon,
after making sure th CI is still green

---
Hannu

On Thu, Oct 8, 2026 at 7:10 AM solai v <solai(dot)cdac(at)gmail(dot)com> wrote:

> Hi Boris,
>
> I reviewed and tested the latest v15 patch for the pgbench faster data
> generation modes.
> My testing results
> - Reproduced the existing behavior before applying the patch using '-I
> dtgG', where both client-side and server-side data generation were
> executed.
> - Applied the v15 patch successfully and built pgbench without errors.
> - Re-ran the same test after applying the patch.
> - After the patch, pgbench reports the warning: 'More than one type of
> data initialization is requested'.
> - Both requested initialization modes were still executed successfully.
> - The test completed successfully without errors.
>
> Is the warning for multiple data-generation modes already covered by
> the existing regression tests?
>
> Based on my testing, the patch works as expected for the tested scenario.
>
> Regards,
> solai
>
>
>