1、新建链接服务器
在图1中选中“链接服务器”，右键选择“新建链接服务器”，如图2，配置相关参数。
2、配置相关参数
在“常规”选项中，填写“链接服务器名称”，访问接口选中
“ Microsoft OLE DB Provider for Oracle”项。其中产品名称、数据源、访问接口字符串请填写配置的Oracle客户端在本地配置的net服务名。如图3所示。
在“安全性”项中，选中使用此安全上下文建立连接，填写远程登录用户和密码，即登录远程Oracle服务器的用户和密码。如图4所示。
3、异构数据库访问语句格式
比如我们建立了链接服务器ORADBCONN，而Oracle中在HRSOFT用户下面建立了表WEBUSER，那么我们的SQL语句就是：
--清空Oracle表中的数据
DELETE FROM ORADBCONN..HRSOFT.WEBUSER
--将SQLServer中的数据写到Oracle中
INSERT into ORADBCONN.. HRSOFT. WEBUSER
SELECT * FROM WEBUSER
如果报告成功，那么我们的数据就已经写入到Oracle中了。
用SELECT * FROM ORADBCONN..HRSOFT.WEBUSER
查看Oracle数据库中是否已经有数据了。
4、链接服务器应用
A、查询Oracle数据表方式一(这种方式，当Oracle与SQLServer的数据类型不一致时经常报错,且速度稍慢)：
select * from [LINK2ORACLE]..[ORACLE_USER_NAME].TABLE_NAME;
我在执行该语句经常报类似错误信息：链接服务器 "LINK2ORACLE" 的 OLE DB 访问接口 "MSDAORA" 为列提供的元数据不一致。对象 ""CMCC"."OS2_GIS_CELL"" 的列 "ISOPENED" (编译时序号为 20)在编译时有 130 的 "DBTYPE"，但在运行时有 5。
B、查询Oracle数据表方式二(经试验，这种方式使用起来很顺畅，不报错，且速度几乎和在Oralce中一样快)：
select * from openquery(LINK2ORACLE,'select * from OracleUserName.TableName')
您可以把openquery()当成表来使用。
C、举个例子（将Oralce用户HRSOFT下的用户表WEBUSER导入到SQLServer2005数据库中）：
select * into OS_GIS_WEBUSER from openquery(LINK2ORACLE,'select * from HRSOFT. WEBUSER)
D、更便捷的方式：通过创建同义词进行便捷查询:
CREATE SYNONYM OS_GIS_CELL FOR [ORACLELK]..[CMCC].OS_GIS_CELL;
select * from os_gis_cell;
select * from os_gis_cell a where a.CellName is null;
注意：涉及 Oracle 部分的 SQL 语句，尤其是 [ORACLELINK]..[ORACLE_USER_NAME].TABLE_NAME 一定要大写，否则会报类似错误：
消息 7314，级别 16，状态 1，第 1 行
链接服务器 "ORACLELK" 的 OLE DB 访问接口 "MSDAORA" 不包含表 ""CMCC"."OS2_gis_CELL""。该表不存在，或者当前用户没有访问该表的权限。
配置數據源：
工程INPLAN:
(DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=192.168.2.6)(PORT=1521)))(CONNECT_DATA=(SID=sa)(SERVER=DEDICATED)))
ERP:
(DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL = TCP)(HOST = 192.168.3.25)(PORT = 1528)))(CONNECT_DATA =(SID= TJTEST)))
(DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL = TCP)(HOST = 192.168.3.25)(PORT = 1528)))(CONNECT_DATA =(SID= TJTEST)))
新建作業里的代碼：
delete from XBLD_Information
INSERT INTO XBLD_Information(job_name, num_layers, JB_PP1, JB_PC1, JB_PP2, JB_PC2, JB_PP3, JB_PC3, PP_PP1, PP_PC1, PP_PP2, PP_PC2, PP_PP3, PP_PC3, PP_PP4,PP_PC4,BM)
select * from openquery(INPLAN,'
select mjs15.job_name, mjs15.num_layers,mjs1.MM1 JB_PP1,mjs1.MC1 JB_PC1,mjs2.MM2 JB_PP2,mjs2.MC2 JB_PC2,mjs3.MM3 JB_PP3,mjs3.MC3 JB_PC3,mjs11.MM1 PP_PP1,mjs11.MC1 PP_PC1,mjs12.MM2 PP_PP2,mjs12.MC2 PP_PC2,mjs13.MM3 PP_PP3,mjs13.MC3 PP_PC3,mjs14.MM4 PP_PP4,mjs14.MC4 PP_PC4,Substr(mjs15.job_name,5,2) BM
from
(select mjl.job_name,
mjl.num_pcbs,
mjl.num_arrays,
mjl.num_panles_,
mjl.num_layers,
substr(mjl.op_size_y_string_,instr(mjl.op_size_y_string_,''='')+1,length(mjl.op_size_y_string_)-instr(mjl.op_size_y_string_,''='')+1) y_size, --祇へ糴
substr(mjl.op_size_x_string_,instr(mjl.op_size_x_string_,''='')+1,length(mjl.op_size_x_string_)-instr(mjl.op_size_x_string_,''='')+1) x_size --祇へ
from tj.my_job_list mjl
) mjs15,
(select *
from
(select mjs.job_name,
decode(DENSE_RANK() over( partition by job_name order by material_name,material_count),1,material_name) MM1,
decode(DENSE_RANK() over( partition by job_name order by material_name,material_count),1,material_count) MC1
from tj.my_job_stackup_material_count mjs
where mjs.type_T=''Core''
)where MM1 is not null
) mjs1,
(select *
from
(select mjs.job_name,
decode(DENSE_RANK() over( partition by job_name order by material_name,material_count),2,material_name) MM2,
decode(DENSE_RANK() over( partition by job_name order by material_name,material_count),2,material_count)MC2
from tj.my_job_stackup_material_count mjs
where mjs.type_T=''Core''
) where MM2 is not null
) mjs2,
(select *
from
(select mjs.job_name,
decode(DENSE_RANK() over( partition by job_name order by material_name,material_count),3,material_name) MM3,
decode(DENSE_RANK() over( partition by job_name order by material_name,material_count),3,material_count)MC3
from tj.my_job_stackup_material_count mjs
where mjs.type_T=''Core''
) where MM3 is not null
) mjs3,
(select *
from
( select mjs.job_name,
decode(DENSE_RANK() over( partition by job_name order by material_name,material_count),1,material_name) MM1,
decode(DENSE_RANK() over( partition by job_name order by material_name,material_count),1,material_count) MC1
from tj.my_job_stackup_material_count mjs
where mjs.type_T=''Prepreg''
)where MM1 is not null
) mjs11,
(select *
from (
select mjs.job_name,
decode(DENSE_RANK() over( partition by job_name order by material_name,material_count),2,material_name) MM2,
decode(DENSE_RANK() over( partition by job_name order by material_name,material_count),2,material_count)MC2
from tj.my_job_stackup_material_count mjs
where mjs.type_T=''Prepreg''
)where MM2 is not null
) mjs12,
(select *
from
( select mjs.job_name,
decode(DENSE_RANK() over( partition by job_name order by material_name,material_count),3,material_name) MM3,
decode(DENSE_RANK() over( partition by job_name order by material_name,material_count),3,material_count)MC3
from tj.my_job_stackup_material_count mjs
where mjs.type_T=''Prepreg''
)
where MM3 is not null
) mjs13,
(select distinct *
from
( select mjs.job_name,
decode(DENSE_RANK() over( partition by job_name order by material_name,material_count),4,material_name) MM4,
decode(DENSE_RANK() over( partition by job_name order by material_name,material_count),4,material_count)MC4
from tj.my_job_stackup_material_count mjs
where mjs.type_T=''Prepreg''
) where MM4 is not null
) mjs14
where mjs15.job_name=mjs1.job_name(+)
and mjs15.job_name=mjs2.job_name(+)
and mjs15.job_name=mjs3.job_name(+)
and mjs15.job_name=mjs11.job_name(+)
and mjs15.job_name=mjs12.job_name(+)
and mjs15.job_name=mjs13.job_name(+)
and mjs15.job_name=mjs14.job_name(+)
')
UPDATE XBLD_Information SET BM=b.Describe from XBLD_Information a, XBLD_BM b where a.bm=b.id
【通过创建SQLServer 2005到 Oracle10g 的链接服务器实现异构数据库数据转换方案】相关文章：
★ 简单实用SQL脚本Part SQLServer 2005 链接服务器
★ SQL Server 2008 维护计划实现数据库备份心得
★ SQLServer 2008 R2中使用Cross apply统计最新数据和最近数据
★ 更改SQL Server 2005数据库中tempdb位置的方法
★ SQL Server SQL Agent服务使用教程小结