Import Oracle Table - Lengths multiplied by 3

clarcombe · Post by **clarcombe** » Thu Mar 28, 2013 4:38 am

I have imported several tables from Oracle (Siebel). When I check the lengths of the fields, they have been multiplied by 3

However, from an Oracle non Siebel schema, I don't get these issues.

I was wondering what I need to look at in order to find out how to avoid this.

chulett · Post by **chulett** » Thu Mar 28, 2013 6:34 am

How did you import the metadata, using which tool/option? Do you have NLS enabled? I've seen this before and I'm sure it has been posted here as well but I don't recall off the top of my head why that happens but I'm sure it's NLS related.

These 'certain Oracle instances' - how does their NLS_LANG setting compare to the others where this doesn't happen?

clarcombe · Post by **clarcombe** » Thu Mar 28, 2013 7:08 am

I imported this via the Import Plug-in Meta Data ORAOCI9. We don't have NLS enabled.

As for the NLS_LANG, can different databases have different NLS_LANG values specified ? I thought it was done at a machine level.

chulett · Post by **chulett** » Thu Mar 28, 2013 7:14 am

clarcombe wrote:As for the NLS_LANG, can different databases have different NLS_LANG values specified ?

Instance level.

clarcombe · Post by **clarcombe** » Thu Mar 28, 2013 7:37 am

I checked with the DBA. For the Siebel databases, they are configured in Unicode whereas the other non-Siebel aren't.

From memory, altering the NLS_LANG wouldn't actually have any effect on the database.

I'll do a search and see if there are Oracle/Unicode workarounds (unless you have any other ideas ?)

chulett · Post by **chulett** » Thu Mar 28, 2013 10:08 am

I'm not sure if there is anything you can do about it other than manually edit it afterwords. You may find that you need to leave it that way when dealing with multi-byte charactersets, however. If your instances use "byte semantics" then this may be a needed workaround. Specifically, when you have a VARCHAR2(10) field, is it sized for 10 characters or 10 bytes? Byte semantics seem to be the default in the single-byte characterset world for some reason.