Folks--
Thanks for the ideas. But bisection just seemed too cumbersome.
In the end I decided to write a data filter in perl which checks
all the data for valid types before putting it into the DB!
Mike
Ed Loehr wrote:
>
> I have found that judicious placement of a few queries (selects, intentional
> errors, etc.) within a long sequence of inserts will help segment them for
> identification of offending lines. Hokie, but it helps me.
...
> Thomas Reinke wrote:
>
> > I've run into this a few times as well. My strategy to "hunt" down
> > the offending line has been to do a "bisection" algorithm.
> >
> > Jose Soares wrote:
> > >
> > > This is also my problem. I'm getting '@!?àù§èé+*_|!&/%§¸' to load a
> > > table with more than 23,000 rows
> > > because I don't know in which line I have to look for the the error.
