本文探讨了Informatica在使用过程中可能出现的问题及其解决方案，包括session错误日志未生成的原因及排查方法、mapping出错时的定位技巧、针对不同目标表类型的加载方式选择建议、批量加载Oracle索引表的最佳实践等。
1.当Informatica的session的错误日志没有产生的可能原因有：
(1).Log directory full
(2).session is invalid
(3).execute permition on the database connection may need to be granted
2.当mapping出错的时候有以下几个定位错误的方法：
(1).Change target writer option from relational to flat file to better identify the problem.
(2).Use the debugger
(3).Use the 'create and add target' option to see output at that point in the mapping.(You can turn any transformation into a target and use the flat file writer to build the output of that transformation)
3.The load type of the target should be set to normal if the target table has keys,indexes defined on it at the physical database. This supplys to oracle tables with keys only.
4.To bulk loading oracle indexed tables choose non-parallel mode and disable the 'enable parallel mode' option in the connection or drop suspend indexes and rebuild them after the bulk loading is complete possiblely in pre-and-post SQL or using an stored procedure transformation.
5.If you configure the session for roe error logging the Integration Services will write row errors to the error log instead of the session log. If you want the Integration Services to write the drop rows to the session log also can configure the session for both data tracing
(1).Log directory full
(2).session is invalid
(3).execute permition on the database connection may need to be granted
2.当mapping出错的时候有以下几个定位错误的方法：
(1).Change target writer option from relational to flat file to better identify the problem.
(2).Use the debugger
(3).Use the 'create and add target' option to see output at that point in the mapping.(You can turn any transformation into a target and use the flat file writer to build the output of that transformation)
3.The load type of the target should be set to normal if the target table has keys,indexes defined on it at the physical database. This supplys to oracle tables with keys only.
4.To bulk loading oracle indexed tables choose non-parallel mode and disable the 'enable parallel mode' option in the connection or drop suspend indexes and rebuild them after the bulk loading is complete possiblely in pre-and-post SQL or using an stored procedure transformation.
5.If you configure the session for roe error logging the Integration Services will write row errors to the error log instead of the session log. If you want the Integration Services to write the drop rows to the session log also can configure the session for both data tracing
扫一扫
1万+
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
