DSXchange

As mentioned you can use Index(), which will give you the starting position of "P/N", you can get the substring of required string from the main string by Input.Field[Index(Input.field,"P/N"),9]. And for the second one, you can get the substring from the first char to the point where you find "P/N" ...

You need to provide space between parameters. You can see the actual value been passed in the director. Are you using the same use id through which you executed in unix? If not check the permissions.

ArndW wrote:How can you set a DB/2 enterprise read to something apart from "sequential" and "parallel"?

Sorry, I dont have access with PX now.

ArndW wrote:Kumar - please tell me what the command line is (apart from the "orchadmin" one I mentioned)? I'm not aware of any other method of doing this so I am curious what option you are referring to.

$dsrecords ds_name
This can be found in UserGuide.pdf, and I dont find it any where else.

It is not a must to sort all the data that passes through the jobs. Sorting data is done only if requried. And in cases if you have filter option need to be held, you can do that before sorting so that sort can have less data to process. If you have any group by operation that need to be done, like ...

Thats good that you have identified the bottle neck. After speaking with your SA, you can share the exact parameters that need to be changed, in your flavour of OS. So that it can be useful for the other users.

Discussion on Hyperion flat file database can be found here.
http://businessintelligence.ittoolbox.c ... ice-487889

This has been discussed many times. One other option, Sort and Enable the KeyChange option. And in transformer, you can track the value of this field and filter based on it.
Or Remove the duplicate, and find a difference with the original file using Difference stage.
And there are many options too.

How do you define "P/N P422" is the required the output. Is it a fixed value for all the set of input srtings. If so you just need to find whether "P/N P422" exist or not. In that case, you can DCount(). If you have another table from which you need to pick up the value you can do lookup and use the...

When I had this problem in 7.0, I was not even able to compile the jobs. You too can check it. If so check the number of process running. You can use ps -eaf | grep dscs and netstat -a | grep ds to find the informations. I was not allowed to restart datastage. But restarting Unix server helped. Then...

Or RoundRobin.

This can be posted in Server Forum.

You can also count the number of records in dataset in command line. Pls search for it.

Make sure your Data fits into memory. Else you may agian stik to Join.

If you search for this string

The number of reject datasets "0"is less than the number of input datasets "1"

. you may find most leading to compiler versioning problem or licensing or not been installed properly.