Re: [cx-oracle-users] Help with fetchraw

[Anthony T. <ant...@gm...>]

On 5/17/05, Chris Dunscombe <cdu...@ya...> wrote:
>=20
> --- Anthony Tuininga <ant...@gm...> wrote:
>=20
> > On 5/17/05, Chris Dunscombe <cdu...@ya...> wrote:
> > >
> > > --- Anthony Tuininga <ant...@gm...> wrote:
> > > > 2) I'm not sure exactly what you are intending with this so I can't
> > > > see why the simple
> > > >
> > > > for row in cursor:
> > > >     process_the_row()
> > > >
> > > > isn't going to work or isn't desirable. Perhaps you could give more
> > > > detail? Thanks.
> > >
> > > The "problem" with:
> > >
> > > for row in cursor:
> > >     process_the_row()
> > >
> > > is that process_the_row() is essentially:
> > >
> > > for col in row:
> > >     rowOut =3D rowOut.append(format_col(col))
> > > rowOutString =3D "".join(rowOut)
> > >
> > > which creates quite an overhead when the tables have lots of columns =
(>40).
> >
> > You should be able to rewrite "process_the_row()" as the following if
> > all you are intending to do is replace None with ""
> >
> > for row in cursor:
> >     rowOutString =3D "".join([v or "" for v in row])
> >
> > if all you are fetching is strings. There is a flag for turning
> > numbers into strings in all cases as well. If you need something more
> > complicated than the above (multiple data types) then I'll need more
> > information. :-)
>=20
> There are multiple datatypes typically varchar,number,date and sometimes =
raw and long.

Ok. So what exactly does the process_row() method do? If you give the
full body I might be able to suggest something that would work better.
Or if you'd rather play with it on your own, that's fine too. :-)