> > > >Adapted Plan
> > > >PLAN JOIN (JOIN (MATRICULA NATURAL, ALUNO INDEX (INTEG_8)),
> > > >PARAM_PER INDEX (INTEG_38))
> > > >
> > > >But the Pk of PARAM_PER is periodoano_paramper, which is
> > > >a Char(6)!
> > >
> > > INTEG doesn't mean integer, it means INTEGrity. What is weird about

this

> > > is that constraint names are showing up in a plan.
> > >
> > > I tested to see whether the new thing in RC3 about named constraints

was

> > > maybe operating on a UNIQUE constraint, causing the constraint name to

be

> > > applied to the enforcing index.
> > >
> > > But, no. The index got a RDB$ name.
> > >
> > > So how does IBExpert manage to pull constraint names into a plan?
> >
> >Most probably by simply doing a string replace for RDB$... names?
> Martijn,
> That wasn't what I was musing on.
>
> I mean - a plan with constraints in the INDEX slots. But maybe that's

what

> the "adapted plan" is meant to show - not meant to be a plan "adapted to
> make it better" or meant to work, but just a tool to show the user where
> the indexes came from?

I guess the latter - instead of showing RDB$... it shows the actual
constraint that defines it. Nice to read. But that's all. Never heard
a user of Database Workbench ask for this...


With regards,

Martijn Tonies
Database Workbench - the developer tool for InterBase & Firebird
Upscene Productions
http://www.upscene.com