DSXchange

The jobs will not fail if you import the tables again. It will prompt you to overwrite or not. It will add new columns and overwrite old ones.

If you output sequential files into one directory then UNIX cat command can combine them. Why would you want to do this?

Do a search this has been covered a lot. Ken Bland posted a very detailed shell script to do exactly what you want.

DSJobReport is a routine supplied with DataStage in the Before/After folders. Here is the long description: After Job subroutine which writes a Job report to a file InputArg is a string which can have 3 ';' seperated fields. Field 1 specifies the report type(0, 1 or 2) as recognized by DSMakeJobRepo...

Sure, switch it to a UV stage. UV stage can access hash files with SQL commands. Do a search. There are lots of good posts on doing this.

Do a search for DS_AUDIT. There are several examples out there already.

Even server jobs are allowed to finish with the exception handler logic in place. I assume PX jobs do the same. The sequence is aborted long before these other jobs are finished. That is important when restarting. You need to check to see if all jobs are finished in that sequence before restarting.

DSStopJob was available in version 5 $INCLUDE DSINCLUDE JOBCONTROL.H RunHandle = DSAttachJob(JobName, DSJ.ERRNONE) JStat = DSStopJob(RunHandle) JStat = DSDetachJob(RunHandle) Here are 3 lines of code to do it. There was lots of code in between to make sure the job was running...

I would use UVwrite. You need the sequence number record as well. I am sure you know UVwrite command.

You need to look at several places. I would start to run top or vmstat during these times. See if you can figure out if you are running out of swap space. In version 5 we had to sleep 20 seconds before we would start a second process. It would then allow us to get more processes running without abor...

Kalyan

I think all of this is fine. I do it. Ray does it. I sure Ken and Craig do it. There is risk involved in going around any product and updating their tables directly. I am saying be safe and not sorry. Backup first.

The problem is there are multiple record types in DS_JOBOBJECTS. So what is field 4 in the ROOT record maybe hash file name in some other record type. So if you create a dictionary item for field 4 and do use OLETYPE in your WHERE clause then you just messed up hash file name. OLETYPE is the record ...

If you do not uderstand Universe then I am not sure I would do this. If you really want to pursue this then short description is field 4 and long description is field 7 of the ROOT record in DS_JOBOBJECTS. You need to be able to create dictionary items for field 4 and 7. Then just update with a WHER...

You are correct. There is no mention of rows counts or run times. Run Reporting Assistant. DSLinks should have most of this.

If you would upgrade then you could use EtlStats. This is a set of jobs and routines to get most of this. It updates a set of tables in the target like ETL_JOB, ETL_ROW_HIST and ETL_LINK. Everything you need except the 2 items below is already built for you. Stage Link Relationship Link Type (input,...

Rows per second are not a good indicator. It should be MB per second. You have 110 columns. Some of those could be large varchars. If you disabled your indexes and not removed them then that will change the performance as well. You always need to ANALYZE TABLE in Oracle when removing indexes or addi...