Re: Patch to show individual statement latencies in pgbench output

Florian Pflug <fgp(at)phlo(dot)org> writes:
> On Aug12, 2010, at 19:48 , Tom Lane wrote:
>> I'm looking through this patch now. It looks mostly good, but I am
>> wondering just exactly what is the rationale for adding comment
>> statements to the data structures, rather than ignoring them as before.

> To be able to include the comments (with an average latency of zero)
> in the latency report. This makes the latency report as
> self-explanatory as the original script was (Think latency report
> copy-and-pasted into an e-mail or wiki). It also has the benefit of
> making the line numbers of the latency report agree to those of the
> original script, which seemed like a natural thing to do, and might
> make some sorts of post-processing easier. It does make doCustom() a
> bit more complex, though.

I had wondered if the original form of the patch printed line numbers
rather than the actual line contents. If that were true then it'd make
sense to include comment lines. In the current form of the patch,
though, I think the output is just as readable without comment lines ---
and I'm not thrilled with having this patch materially affect the
behavior for cases where -r wasn't even specified.

regards, tom lane