> pavel.steh...@gmail.com wrote: > >> b...@yugabyte.com wrote: >> >> As it happens, Oracle's PL/SQL has a "goto" statement. But PL/pgSQL does >> not. (I assume that this is because "goto" is considered a bad thing.) But >> PL/SQL programmers do use it. However, the doc section: > > The reason why PL/pgSQL has not "goto" statement is mainly technological. > PL/pgSQL is an interpreter of high level commands. For this kind of > interpreter the "goto" - unstructured jump cannot be effectively implemented. > PL/pgSQL is very simple, and relatively fast (expressions are slow due > evaluation by SQL executor), but "goto" cannot be implemented there. > Interpreter of PL/pgSQL is very different from the more usual p-code > interpreter.
It’s interesting to know that the reason that PL/pgSQL doesn’t support “goto” is an implementation restriction rather than a purist stance. Thanks! I mentioned PL/SQL only to say that it does not support the premature exit from a block statement that PL/pgSQL _does_ support (and document). I accept, now, that I’ll never know the rationale for this.
