Hello,

I import a lot of data from CSV files into SQL Server 2000 using the
DTS Import/Export Wizard.

I've just discovered a very worrying problem after a user reported that
I wasn't reporting the financials correctly. After looking into the
problem it turns out that the wizard is somehow skipping certain
records and returning a successful message.

I've re-imported the file numerous times, ensured that the source data
is valid and I can't come up with a solution - records with the dates
29-JUN-2005 and 30-JUN-2005 are just being missed.

Has anyone seen this before or have any ideas on possible solutions?

Thanks,

Ryan

Please post back with the data as unmodified as you can including rows
before the failing rows that were successful and the failing ones. It's
probably a formatting issue or a conversion issue. I would have to hope
that an error is happening somewhere, or the data is somehow ending up
somewhere else and it's not just as you say. I've never experienced
something flaky like you're describing.

Ryan wrote:

A very quick test for this would be to take the offending rows out of
the file and try to enter those rows manually into the table in
question. If you cannot do it through Query Analyser then it is
unlikely they will go in through DTS.

It is I would say almost definitely a conversion issue. Would it be
possible to get the structure of the destination table and the offending
rows in a file so we can have a go at reproducing.

Have you tried to not use the wizard and build the package yourself.
You may get more joy with errors this way.

Another thing to try is look at the file in a Hex Editor and make sure
the offending rows are terminated properly

Allan



"Ryan" <ryan.heywood (AT) gmail (DOT) com> wrote