Hi,
I need to make an insert of 1M rows, approximately 1000 of them raise an error.
Any way to isolate these rows?
Thanks
We are looking into adding an output port to the databse writer, but if you can model the validation procedure in KNIME in the mean time, maybe it could be a semi-reasonable workaround. Do you know what is causing the error?
Yes it is a duplication of the key. To avoid this problem i need another reader and a join but take too much time.
I need now to stop the execution or simply alert the user if one row raise an error. (even at the end of the database writer)
Is there a way to do this?
The node does not stop the execution but only print to console the first 10 rows in error therefore i can't use a try-catch. The only idea is parsing the log file looking for the string "Error". Before to continue along this path i prefere to hear a second opinion.
Hi Darbon
I cannot help you but I have had your same problem when inserting duplicate rows in Oracle and DB2 tables.
If anybody had a quick, alternative way to manage the problem, even using java, it would be great!
Nik