Provided: A computer sqlldr import data to a computer using the Oracle batch B. Separate data and queries are normal.
A in the copy source data file to the computer C ctl and the copy Oracle installation file B and C are mounted to the computer.
Computer use C ctl after the discovery of the source file into the native Oracle. The imported data partition errors, such as AAA @ bbb @ ccc data format, when introduced sqlldr separated by @. But the actual results are not introduced column1 = aaa, column2 = bbb, column3 = ccc; some data becomes column1 = aaa @, column2 = bbb @, column3 = '' in this case, and then use the select * from table xxx where column1 = 'keyword' query time, non-Chinese keyword search properly. Chinese keyword can not find the data (actual data). After Baidu Online agreed that caused the server and client character set is inconsistent reasons. Then use the online method to query the Oracle server computer C character set encoding and client character set encoding. Before and after the system variables NLS, NLS registry file and server are set to become the same AMERICAN_AMERICA.AL32UTF8, after importing data and test queries still problems mentioned above. After comparison of the computer system environment and the computer C A variable settings and registry settings. To modify the computer registry NLS C and A as SIMPLIFEID CHINESE_CHINA.AL32UTF8, after test import and queries, problems remain. Then modify the value of the system variable NLS_LANG AMERICAN_AMERICA.ZHS16GBK. After all the normal tests. Determine the cause of the registry is not a problem, the environment variable is the cause of the problem. Why be so specific settings unknown. Left to future research.