Обсуждение: Conditional compilation

Поиск
Список
Период
Сортировка

Conditional compilation

От
Thiemo Kellner
Дата:
Hi

Does PostgreSQL have something like Oracle's conditional compilation? 
This is sort of an if then statement that gets evaluated on 
compilation/installation time of PL/SQL code. If the condition is met, 
the code until the $END gets compiled. It is even possible to switch 
on/off parts of single statements. I suppose it is sort of a 
preprocessor that removes the code part from $IF until $END if the 
condition is not met.

Kind regards

Thiemo




Re: Conditional compilation

От
Ron
Дата:
On 11/12/23 09:32, Thiemo Kellner wrote:
> Hi
>
> Does PostgreSQL have something like Oracle's conditional compilation? This 
> is sort of an if then statement that gets evaluated on 
> compilation/installation time of PL/SQL code. If the condition is met, the 
> code until the $END gets compiled. It is even possible to switch on/off 
> parts of single statements. I suppose it is sort of a preprocessor that 
> removes the code part from $IF until $END if the condition is not met.

Pl/PgSQL is an interpreted language; there is no compilation.  At 
creation/installation, it just (I think) does syntax checks; it definitely 
doesn't care if a table exists or not.

-- 
Born in Arizona, moved to Babylonia.



Re: Conditional compilation

От
Tom Lane
Дата:
Ron <ronljohnsonjr@gmail.com> writes:
> On 11/12/23 09:32, Thiemo Kellner wrote:
>> Does PostgreSQL have something like Oracle's conditional compilation? This 
>> is sort of an if then statement that gets evaluated on 
>> compilation/installation time of PL/SQL code. If the condition is met, the 
>> code until the $END gets compiled. It is even possible to switch on/off 
>> parts of single statements. I suppose it is sort of a preprocessor that 
>> removes the code part from $IF until $END if the condition is not met.

> Pl/PgSQL is an interpreted language; there is no compilation.  At 
> creation/installation, it just (I think) does syntax checks; it definitely 
> doesn't care if a table exists or not.

Yeah.  You can get at least some of the effect of this by just writing
if-statements.  The non-executed chunks of code still have to be
grammatically valid SQL, but they don't ever reach parse analysis
so nothing happens beyond minimal syntax checking.

If that's not enough, what you'll want to look at is using EXECUTE
to execute dynamically-constructed SQL.

            regards, tom lane



Re: Conditional compilation

От
Thiemo Kellner
Дата:
Thanks for pointing out. The only use case of conditional compilation, I 
can think of, is, that one can have arbitrary logging code in some 
environments without performance penalty in others, e.g. log as hell in 
dev, no logging at all in prod without the "function" having to execute 
checks for every and each logging statement put (as the check has been 
done on installation already).

begin
     do_something;
     $if check_if_env_is_dev $then
         do_some_logging;
     $end
     do_more_stuff;
end;

Am 12.11.2023 um 16:58 schrieb Tom Lane:
> Ron <ronljohnsonjr@gmail.com> writes:
>> On 11/12/23 09:32, Thiemo Kellner wrote:
>>> Does PostgreSQL have something like Oracle's conditional compilation? This
>>> is sort of an if then statement that gets evaluated on
>>> compilation/installation time of PL/SQL code. If the condition is met, the
>>> code until the $END gets compiled. It is even possible to switch on/off
>>> parts of single statements. I suppose it is sort of a preprocessor that
>>> removes the code part from $IF until $END if the condition is not met.
>> Pl/PgSQL is an interpreted language; there is no compilation.  At
>> creation/installation, it just (I think) does syntax checks; it definitely
>> doesn't care if a table exists or not.
> Yeah.  You can get at least some of the effect of this by just writing
> if-statements.  The non-executed chunks of code still have to be
> grammatically valid SQL, but they don't ever reach parse analysis
> so nothing happens beyond minimal syntax checking.
>
> If that's not enough, what you'll want to look at is using EXECUTE
> to execute dynamically-constructed SQL.
>
>             regards, tom lane