NLS_LENGTH_SEMANTICS
936749Nov 19 2012 — edited Nov 19 2012Hi,
We're migrating from Western Europe character set to UTF8 and we see that some columns is not "sized" to store the data as some characters now are multi bytes (as expected)
As I see it we basically have 3 options:
1. Do not change character set
2. Increase size of column
3. change NLS_LENGTH_SEMANTICS to char
Alternative 1: is possible but really more of a last way out of this issue.
Alternative 2: is also an option and easy for us DBA's to implement, but it has some big downsides for developers as they may have to change size different places in the application.
Alternative 3: This one I'm a bit unsure of. The default is byte and I guess it's a reason why Oracle has byte's as default? Is there any downside to running char if not why would Oracle use byte and why is there two options?
If someone know of a 4'th and 5'th I'd be grateful for your inputs.
Please share your knowledge :)