We are testing our new 18.104.22.168 database using Oracle Linux 6. We created the database using the AL32UTF8 NLS Character set. We have tried using sqlldr to insert a few records that contain Russian and Chinese characters as a test. We can not seem to get them into the database in the correct format. For example, we can see the correct characters in the file we are trying to load on the Linux server, but once we load them into a table in the database, some of the characters are not displayed correctly (using SQL*Developer to select them out).
We can set the values within a column on the table by inserting them into the table and then select them out and they are correect, so it appears the problem is not in the database, but in the way sqlldr inserts them. We have tried several settings on the Linux server to set the NLS_LANG environment to AMERICAN_AMERICA.AL32UTF8, AMERICAN_AMERICA.UTF8, etc. without success.
Can someone provide us with any guidance on this? Would really appreciate any advice as to what we are not getting here.
The characterset of the database does not change the language used in your input data file. The character set of the datafile can be set up by using the NLS_LANG parameter or by specifying a SQL*Loader CHARACTERSET parameter. I suggest to move this question to the appropriate forum: Export/Import/SQL Loader & External Tables for closer topic alignment.
We are testing our new 22.214.171.124 database using Oracle Linux 6. ...
Pl note that currently only 126.96.36.199 is certified/supported in Linux 6 - http://docs.oracle.com/cd/E11882_01/relnotes.112/e23558/toc.htm#CHDFHIEA - 188.8.131.52 is not - so things may or may not work as advertised.
What is the locale setting in Unix ?
The correct NLS_LANG setting in Unix Environments [ID 264157.1]
## we can see the correct characters in the file we are trying to load on the Linux server
How do you look at them? Do you see both Chinese and Russian correctly in a single file? Note, Russian is (partially) supported by Chinese character sets so the fact that you see both languages at once in a file does not necessarily mean that the file is encoded in UTF-8.