On Wed, Jul 14, 2021 at 02:02:18PM -0400, Tom Lane wrote:
>
> Hm. I guess this point (i.e. that the Param substitution should result
> in the identical plan tree as writing a literal constant would have)
> has some force. I still feel like your application is pretty shaky,
> but I don't really have any ideas about feasible ways to make it better.
I totally agree that it's quite shaky, but I also don't have a better idea on
how to improve it. In the meantime it already saved my former colleagues and I
(and I'm assuming other people) many hours of boring and tedious work.
> Will push the patch.
Thanks a lot!