Page 3 of 3
Re: Copy Rejected Data
Posted: Tue Sep 23, 2014 9:41 pm
by ClevelandSteve
knicely87 wrote:Since you are using a shared directory maybe you need to have unique file names for each node?
Example:
Copy myTable From '/myDirectory/*' On Any Node
Rejected Data '/rejects_directory/node1.rej' On node_1, '/rejects_directory/node2.rej' on node_2
Exceptions '/exceptions_directory/node1.exp' On node_1, '/exceptions_directory/node2.exp' on node_2
Well the files I'm loading are numbered, i.e. source001, source002, etc. Thus my copy statement reads
Code: Select all
Copy myTable From '/myDirectory/source*' On Any Node
Therefore the reject file names would have unique file names, i.e. source001-Copy-From-Rejected, etc.
Re: Copy Rejected Data
Posted: Tue Sep 23, 2014 9:47 pm
by ClevelandSteve
scutter wrote:I didn’t see any information in this thread on the version being used - or maybe I missed it. If you’re using 7.0 or later you can use REJECTED DATA AS TABLE and avoid this issue. But I imagine you would have tried that already so you must be on 6.1.3?
Yes it's version 6.1.3. I opened a ticket with Vertica support on this and went in circles with them for weeks. In the end they basically said upgrade to 7 and use REJECTED DATA AS TABLE. Unfortunately upgrading the database would fall under the responsibilities of our SysOps group and it would never be a high enough priority for them to do so.
Re: Copy Rejected Data
Posted: Wed Sep 24, 2014 2:43 pm
by scutter
it would never be a high enough priority for them to do so.
Just find a way to crash the database regularly so that they get paged often in the middle of the night. Then they’ll be motivated to do an upgrade
Re: Copy Rejected Data
Posted: Mon Dec 01, 2014 4:04 pm
by Benjamin_e
In order toadd rejections and exceptions you must specified directories that already exsits . The copy not create the folders for you.. for bulk upload.