Overflow error

Hi:

I am trying to pump data from Sybase to SQL Server using SSIS and I get this error:

Conversion failed because the data overflowed the specified type

The data on the external column metadata shows as type database timestamp, as does the output column. The database values are all datetime, coming in through OLEDB to Sybase. Any idea what could be going on here?

Thanks,

Kayda

I should correct, the data is coming from a datetime field in Sybase to a datetime field in SQL Server. The values in Sybase are all datetime--it runs for awhile and then gives this error.

Any ideas?
|||Just wondering whether anyone has the answer to this--this problem is cropping up again. There isn't nothing wrong in Sybase with the datetime field as far as I can tell...|||

A quick search did not yield any good results on differences between SQL Server and Sybase DATETIME data types. The thing that seems most likely to me that there are different data ranges supported by each platform.

Can you select a MIN and MAX of the problem field when the error is occuring?

|||

Yes, I looked for NULLs and did a MIN and MAX on Sybase--just normal dates from 2004 up until "yesterday". I am running this pump on many other similar tables with the same time field. Is there a way to figure out exactly is going on?

Thanks,

Kayda

|||

Hi Kayda,

Is the error coming from the source component, or the destination (on insertion)?

What are your source and destination components?

Can you try to hook up an error output, and have the error rows redirected to see if you can identify the data that is failing?

Thanks

Mark

|||

I tried something like that, but didn't seem to work. Do you mean just hooking a table to the red arrow or putting a transform in between? Could you give me an idea of what might work in this particular case?

|||

Kayda_SQL wrote:

I tried something like that, but didn't seem to work. Do you mean just hooking a table to the red arrow or putting a transform in between? Could you give me an idea of what might work in this particular case?

It might be faster to connect the red arrow to a row count transformation and then add a data viewer to that path. But yes, you could hook that red arrow up to an OLE DB destination and store the contents in a table.

Just make sure that you aren't using the fast load option (for this test) so that you get each row in error instead of a full batch. Know that this is much slower though -- but it should give you the data at least along with an error code.