DSXchange

You need to telnet into your DataStage server. If you are on Windows then you will get a TCL prompt ">". LOGTO YourProject PTERM CASE NOINVERT (paste SQL statement above.) If on UNIX: cd `cat /.dshome` ./bin/uv (now you should see the TCL prompt ">") LOGTO YourProject PTERM...

You can add columns if you double click on either the imported table definition or the saved one. So save it to a different name and edit it.

Ray

How do they get the metadata into Redbrick? I was assuming they could get it out of the MetaStage respository easier than Universe.

Why not MetaStage?

You could do this in a shell or perl script. You could uv, uvsh, ds or dssh to execute SQL to retrieve information. You could also use dsjob.

What exactly do you want?

I did not think it is possible but it will allow you to have a transform without primary input. You need stage variables. Like x=1 initial value then x=x+1. Next you need a constrait like x<100. Try it. I have seen people do this to create time and day dimensions. Saves coding the whole thing in a b...

You have to go to the Windows server. They are all normal DSNs setup usually with %SystemRoot%\system32\odbcad32.exe in the control panel.

Try these 2 posts.

http://www.dsxchange.com/viewtopic.php?t=84960
http://www.dsxchange.com/viewtopic.php?t=90300

The cache limits are preloaded in uvconfig. You have to stop and start DataStage to change them. You cannot set these to arbitrary values either. There is a command called shmtest or something close to that which will tell you the max values you can set them at. If you exceed these limits then DataS...

They have disabled it except for internal use. If you use localuv as a DSN then it works for UV stages and internally in BASIC. Otherwise forget it.

I would dump the first few rows to a sequential file just to make sure the date is formated the way you think it is. Depending on the version you may need milliseconds.

Ray

I think this question is valid. MetaStage is really designed to extract meta data from DataStage, Oracle, SQL Server and other databases, ERwin, BO, Cognos and other reporting tools. Can it get meta data from other sources like spreadsheets?

You might be able to do this better in one job. Ask Ray.

If you cannot get it this way then try this SQL at TCL. SELECT DS_JOBOBJECTS.OLETYPE as LinkType, DS_JOBS.NAME as JobName, DS_JOBOBJECTS.NAME as LinkName, DS_JOBS.CATEGORY as Category FROM DS_JOBOBJECTS, DS_JOBS WHERE OBJIDNO = JOBNO Group by LinkType, JobName, LinkName, Category ; Let us know if it...

I am not sure you can do this. Each job runs independently. Unless you have only one record in a file and then run all the jobs then process the next file with only one record in it.

That sounds really ugly. Explain why you think you need to do this.