Re: UPDATE with data at exec and CURRENT OF question

["Hiroshi Saito" <z-saito(at)guitar(dot)ocn(dot)ne(dot)jp>]

Hi.

From: "Lothar Behrens" <lothar(dot)behrens(at)lollisoft(dot)de>

> "Hiroshi Saito" schrieb:
> > Hi.
> >
> > ----- Original Message -----
> > From: <lothar(dot)behrens(at)lollisoft(dot)de>
> >
> > >
> > > "Hiroshi Saito" schrieb:
> > >> Hi.
> > >>
> > >> Ah thanks, It is the environment which can reproduce me.
> > >> Then, I am beginning to make a test code.  There is likely to be a problem
> > >> of some except for SQL_FETCH_NEXT....
> > >
> > > I have created a small test application only with plain ODBC stuff. I
> > > could reproduce
> > > my problem also on Windows with MS SQL Server 2000 !
> > >
> > > You could grab the code here:
> > >
> > > http://lbdmf.cvs.sourceforge.net/lbdmf/CPP/RegressionTests/Database/Plain-ODBC-Test-Blob/test.cpp?revision=1.4&view=markup
> > >
> > > The relevant function is testBlobUpdate.
> > >
> > > On PostgreSQL I did not yet have tested on actual drivers, but when
> > > the MS SQL driver behaves similar, I suppose still an error in my
> > > code.
> > >
> > > Regards, Lothar
> >
> > Uga...Sorry, SQL_CURSOR_KEYSET_DRIVEN cannot be supported yet.:-(
> > Although I tried it as remembered, it is very difficult correction....
> > As for the present version, SQL_CURSOR_FORWARD_ONLY and
> > SQL_CURSOR_STATIC are supported.
> > To be sure, it needs to exist as TODO.!
>
> I have solved it by binding the binary column as usual. Then I have
> set the value of
> the column to an unique value - maybe cursorname - and have updated
> the database
> with that value set.
>
> Then I was able to use a different query (prepared) with a simple
> where clause and this was successfull.
>
> So this may be a solution for the TODO ?
>
> I have figured out some problems doing this on MS SQL but this another
> issue and propably
> only a cast for the where clause because they do strikter type
> checking.

Oh, Cool, It is great solution. I'm sorry in explanation having been lacking.
Sorry, my explanation is insufficient...SQL_CURSOR_KEYSET_DRIVEN
does not become effective together with DECLARE/FETCH options.
It can be used as the option of UPDATE cursor.
Correctly, it should have made it TODO to use with it.

Anyway, I'm sorry confused the user.

Thanks!

Regards,
Hiroshi Saito