ORA-00096: invalid value for parameter plsql_compiler_flags, must be from among NON_DEBUG, DEBUG, INTERPRETED, NATIVE
author：skate
time：2009/10/16
现象：
协助技术搭建一个测试库
源库：oracle11g
目标库：oracle9.2.0.4
考虑imp/exp版本兼容的问题，我用9i的client客户端exp 11g数据，然后再用9i的client客户端把数据imp进9i数据库，
结果就出现如下错误
[oracle@20081006-1724 ~]$ imp userid=sys/aibo fromuser=skate touser=skate file=/tmp/hpo520315.dmp log=/tmp/hpo5203152.log buffer=409600000 ignore=y
Import: Release 9.2.0.4.0 - Production on Fri Oct 16 12:09:01 2009
Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.
IMP-00058: ORACLE error 28009 encountered
ORA-28009: connection to sys should be as sysdba or sysoperUsername: sys as sysdba
Password:
Connected to: Oracle9i Enterprise Edition Release 9.2.0.4.0 - 64bit Production
With the Partitioning, OLAP and Oracle Data Mining options
JServer Release 9.2.0.4.0 - Production
Export file created by EXPORT:V09.02.00 via conventional path
import done in AL32UTF8 character set and AL16UTF16 NCHAR character set
IMP-00003: ORACLE error 96 encountered
ORA-00096: invalid value for parameter plsql_compiler_flags, must be from among NON_DEBUG, DEBUG, INTERPRETED, NATIVE
IMP-00000: Import terminated unsuccessfully
解决过程：
在网上google，发现如下方法
Solution Description --------------------
You need to set the plsql_compiler_flags parameter to another value then NATIVE before exporting your data.
So, either dynamically update this parameter with an 'alter system set' before exporting any data, for example:
SQL> alter system set plsql_compiler_flags=INTERPRETED,NON_DEBUG;
or update your init.ora to comment out this parameter (the default 'INTERPRETED','NON_DEBUG' will then be used)
or set it to any other possible values, such 'INTERPRETED'. If you change your init.ora, you need to restart your
database.
Explanation ----------- You are encountering a bug that will be fixed in the upcoming patchsets. plsql_compiler_flags
determines whether PL/SQL code is compiled native or interpreted and if debug information is included. Possible values
are:
'INTERPRETED' - compile in interpreted mode
'NATIVE' - compile in native mode
'DEBUG' - include debug information
'NON_DEBUG' - no debug information
Defaults are: 'INTERPRETED','NON_DEBUG'
但是无法解决，最后我把源库的数据库对象导出来为sql文件，然后再把sql文件在目标库上执行，最后通过Dblink完成字典表的初始化。
到此测试库搭建起来了，但编译数据库对象是发现N多对象没有编译通过，提示plsql_compiler_flags参数值不正确，想到刚才调整了这
个参数，于是查看下这个参数的当前值
SQL> show parameter plsql_compiler_flags
发现plsql_compiler_flags的值为”NATIVE“
于是更改参数plsql_compiler_flags为默认值，如下：
SQL> alter system set plsql_compiler_flags=INTERPRETED,NON_DEBUG;
结果在编译，都通过了
最后配置了分布式数据库的链接关系，至此一套测试环境搭建ok
考虑中：
我们目前总共需要三套测试，开发环境，要保证其中的版本一致问题，一直是我琢磨的问题，怎样做才可以把我解脱出来，只需要某人
运行一个命令，就可以保证版本一致
-------end-----