DSXchange

I think because the filename is included in the output. The filenames are different.

I need to know which of these files is important. I think we need a FAQ of which files are critical. I have a couple shell scripts I can post tomorrow which backup some of these to the home directory of dsadm. Maybe that will help.

On some of these files I made updatable by dsadm only. This protects from new DSNs and other changes going into effect without admin knowing about it. They will always need read permission.

It is a lot easier to read and write to hash files or sequential files in BASIC. Stick to those. Write a job to extract the data to a hash file. You could easily process this in BASIC use Ray's routine.

In the BASIC.pdf it talks about it. Are you talking about SQL against hash files or Universe files or other databases. Most of the time this should be done in jobs. You lose the GUI viewing capabilities when you run SQL in routines. It sort of defeats the purpose of DataStage. Maybe you could give u...

I think some of this has been posted in the past. Some of this is a personal preference. You need to develop your own naming conventions. I think Ray or Ken may charge you a lot of money to give you their standards. Many years work went into developing these documents. There was an old newsletter on...

That looks like part of compile routine I posted. It is incomplete. If you try to compile jobs that are running then you can have problems. If you clear RT_STATUS files while jobs are running then this also is not good. The jobs continue to run in the background and are no longer talking to the Data...

This has been covered a lot. The field you want is called SEVERITY. Do a search for RT_LOG.

I have used older versions of MKS with DataStage. It works really well. I have used it for many years. I like running Korn shell scripts in DOS.

If you go to my tips page and download EtlStats.zip. It has a job called EtlChkSeqJobsNotRun. It will find all the jobs not run since the sequence is last started. It will email you the log files of all the jobs that did not run as html pages attached to the email message. I still agree with Ray. Yo...

It is in the zip. You can change the path to wherever you put this file. You need to be able to run the sequence SeqPrepareEtlStats before all of this is ready to add to each job. The email is controlled by a shell script called Email.ksh. It should be in a folder called Scripts below each project y...

I posted a routine called GetRowLength. It is part of EtlStats. I have added it to my jobs so I can automatically get MBs/second but ti should not be hard. There is a job called LoadEtlJobLink that will need to call this routine and add a column called AVE_ROW_LENGTH and update it with the value ret...

You need to download EtlStats.zip from my tips page. It is a series of jobs which run dsjob -report XML and load the data into Oracle tables. The create table scripts are all there. There are lots of reports to give you row counts by sequence or run times of all sequences. There is a way to automati...

On 3. you can download DSaveAsBmp.bat on my tips page. It will run the command on every job in a project. It is a DOS batch file. It runs dsjob to get a list of jobs then runs the Designer.exe on each job in the project. Saves lots of time. This files can be used by the GenHtml jobs also on my tips ...

To get row counts in XML format then

../../DSEngine/bin/dsjob -report #projName# #jobName# XML >#jobName#.xml

There is another form of this report that is not XML if you leave off "XML" then you get that version.