Hello Tom,
>> pgbench (well, at least if I succeed in getting boolean expressions and
>> setting variables, which is just a maybe), but this kind of if in the
>> middle of expression does not make much sense for a pgbench script where
>> "if" must be evaluated at execution time, not parse time.
>
> Well, it's not really clear to me why that would be true.
For example, how can you PREPARE a possibly combinatorial thing?
SELECT \if ... XX \else YY \endif
FROM \if ... ZZ \else WW \endif
WHERE \if ... AA \else BB \endif ;
Or the kind of operation:
\if ... SELECT * \else DELETE \endif FROM table WHERE condition;
Even the structure can be changed somehow:
SELECT \if ... 1 ; SELECT 2 \endif ;
> If it actually is impossible to give pgbench equivalent behavior, we'll
> just have to live with the discrepancy,
Yep.
> but ISTM it could probably be made to work.
Even if it could somehow, I do not see it as a useful feature for pgbench.
I also lack a good use case for psql for this feature.
--
Fabien.