on 09-05-2013 3:15 PM
Hello,
How to prevent data duplicating when Job get failed in Data Services.
By using table comparison, Auto correct load we can avoid data duplicating but it would be performance issue.
What are the possible ways are existed to handle such kind of scenarios.
Santhosh
Hi Santhosh,
Use try catch functionality and inside catch write a sql statement to delete data for current load.
I hope this will helpfull for you.
Regards
Asgar
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi,
Possible options
1. Add a script to the dataflow that prevents loading duplicate records while re excuting a job.
2. Use option as "delete data before loading option".
3. Identify the staging tables,delete the data in that tables and re run the job.
But we are using first option in my current project.
Regards,
Narasimha.
Adding to the above points, there are some more methods including:
Data flow design
You can design the replication job to completely replace the target table each time the job is executed. You might choose to use this method if the number of rows changed in the target table is a significant percentage of the target table. You can also use techniques such as bulk loading options to improve the overall performance of the replication job
Pre-load SQL commands
You can use pre-load SQL commands to remove partial database updates that have occurred during
an incomplete job execution. Typically, the pre-load SQL command deletes rows based on a variable
set before the partial insertion began.
Thanks.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi,
Kindly refer the chapter 19 "Recovery Mechanisms" in Designer guide 'sbo411_ds_designer_en' (page no: 645).
Thanks,
Parthiban
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
95 | |
11 | |
10 | |
9 | |
9 | |
7 | |
6 | |
5 | |
5 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.