从oracle插入数据时，基于磁盘的内存OLTP的性能改进 [英] Performance Improvement of In memory OLTP over disk based table while inserting data from oracle

问题描述

我正在针对基于磁盘的表检查插入时序到内存表的性能。这里有详细信息

I was checking the performance of insert timing into In-memory table against disk based table. here are the details

使用SSIS我从Oracle视图中提取数据，该视图有大约800,000条记录，大约需要03:30分钟加载到内存表中，大约需要相同的时间插入基于磁盘的表。

Using SSIS I am pulling data from Oracle view which has about 800,000 records which takes about 03:30 minutes while loading into in-memory table and it takes about same time while inserting into disk based table.

任何人都可以让我知道我在这里错过了什么让它运行得更快。从源到目标只有1：1。

Can anybody please let me know what am I missing here to make it run faster. It's just 1:1 from source to target.

谢谢

 

推荐答案

如果您有足够的RAM并已分配到SQL（最大服务器内存）然后，即使在基于磁盘的表中，所有处理也可以在插入操作进行时进入内存。在这种情况下，使用内存表几乎没有或没有性能优势
。

If you have enough RAM onboard and allocated to SQL (max server memory) then all the processing may be brought into memory anyway while insert operation is going on, even in disk-based table. In that case, there would be little or no performance benefit using in-memory table.

只是推测。除此之外，我们需要更多细节才能回答您的问题。例如目标表的DDL，Oracle客户端/驱动程序版本（32位或64位）SSIS包详细信息等。

Just speculating. Beyond that, we would need more details in order to answer your question. Such as DDL of target table, Oracle client/driver version (32-bit or 64-bit) SSIS package details, etc.

如果您还没有，则应考虑实施< a href ="https://docs.microsoft.com/en-us/sql/integration-services/data-flow/ole-db-destination?view=sql-server-2017#fast-load-options">
快速加载选项如果你想提高速度。 

If you're not already, you should consider implementing fast load option if you want to improve speed. 

另外，考虑一下Attunity Oracle连接器，以进一步提高加载速度。

Also, consider Attunity Oracle connector, to increase load speed further.

HTH，

这篇关于从oracle插入数据时，基于磁盘的内存OLTP的性能改进的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！