Wow thanks to all that replied - you folks are correct. The "_p" and the
others are vars.
_p RECORD;
I won't forget that one for a long time (however, I do drink :-))
Johnf
On Friday 09 April 2010 10:32:51 am Pavel Stehule wrote:
> Hello
>
> 2010/4/9 John <johnf@jfcomputer.com>:
> > Hi,
> > I am reviewing a function written by some xTuple guys. What is
> > interesting about it is it uses the "INTO" statement like
> >
> > select something into _p from sometable where somecriteria.
> >
> > The function contiunes and uses the data retreived
> > _p.somefield_name
> >
> > And then the function ends.
> >
> >
> > Ok my question:
> >
> > I also thought the select "into" created a real table. But after running
> > the function the table does not exist. I see no where that a 'drop' is
> > issued. In fact the function uses lot's of select into's like (_test, _r,
> > etc..). So would some kind soul explain what is happening.
>
> _p is record variable. See some lines before. There will be DECLARE part
>
> DECLARE p RECORD;
>
> There is plpgsql's SELECT INTO and SQL's SELECT INTO with little bit
> different syntax. First - target is list of variables or record
> variable, second - target is table.
>
> > Could it be that "_p" is drop automaticly when the function ends?
> > Something to do with scope.
>
> _p is just variable
>
> regards
> Pavel Stehule
>
> > Could it have something to do with the fact the function returns only an
> > integer? And that causes the table to be drop.
> >
> > As you can see I'm lost here!
> >
> >
> > Johnf