Hello,

I have created one package which will use Merge Join transformation (Inner Join) to relate two table to get some result.

The both table contain 30,00,000 records. The package is taking much time to execute.(in merge join transformation).

Is there any alternative solution to handle this problem,spacialy if there is huge data to join or how to increase the package performance.   

30 millions of rows? Do the tables have appropriate indexes? 

http://beyondrelational.com/modules/2/blogs/101/posts/13416/ssis-97-when-merge-join-is-your-friend.aspx

Hi Vinay,

What is the use of merge join transformation  in your package ? If i understand you are comparing two tables from sql Server Then why don't you MERGE in sql in execute sql task ? Merge join transformation is a  Partially Blocking transformation  and it impcats the performance to a great extent.

Refer

http://www.bimonkey.com/2010/10/the-merge-join-transformation/

Thanks for the reply,

My one source is Oracle (TBL1)and another source is SQL server table(TBL2). One column is same in both the table

I want all the columns of the TBL1 and Primary key column of TBL2 after applying INNER join ON both the tables (As a output of Merge Join).

In this case i am not able to use Execute SQL task.

Is there any alternate way to achieve this task ?  

Hello ,

Is there any suggestion on this issue

Did you try to partition both tables on the columns you want to join and do a hash join?