DSXchange

My logic is based on your last column. I hope you really do have that column and didnt just make that up for our clarification.

You will be needing 5 stage variables. Something like Cond1 | if (In.Col4 = 'Start') then In.Col2 else (if In.Col2 <> PrevCol2 then Cond1:',':In.Col2 else Cond1) Cond2 | if (In.Col4 = 'Start') then In.Col3 else (if In.Col3 <> PrevCol3 then Cond2:',':In.Col3 else Cond2)...

Yes, you are headed in the right direction. Use 'Start' as your refresh point to clear out the concatenation of present to the previous. Basically you have to check if present row is different then the previous, if it is concatenate them together. Do it for botht he columns and mark the row right be...

Let DBA/ETL group be the owner of the application. Other application areas can get this groups resources to get their ETL work done. This way the control remains within one group. And that group will contain junior to architecture level resources who are comfortable with the tool.

Read postby Vincent. It will help you use @INROWNUM along with @PARTITIONNUM and @NUMPARTITIONS.

I take my words back. UtilityAbortToLog wont work inside the px transformer. The only other work around that i can think of at the moment is writing an after job subroutine and check for number of rows in that particular link. If its more than what you want, you can call DSLogFatal() routine or even...

How often do you pass a commit? And also are you performing and deletes before doing your inserts? What database is it?

WHat is 'PUSH_INFOPACKAGE_NAME' in your job? Is that the only error message you get? Please provide any surrounding warning messages.

Used it in a server job. Can test it on px tomorrow and let you know. It really wont bite, you can try it out for yourself. Fix up a small test job and test it out.

What about UtilityAbortToLog. That will work.

You cannot do that while the link is open and data is being processed. You need to do that in the after job routine or something.

You need a work around. Either load your file to a temp table and then perform regular sql. OR create a dummy key with running sequential number while creating the hashed file so that each and every record is retained. Access that hashed file via Universe Stage and them do sql join at the expense of...

No. You can even get rid of the user variable stage. Connect the execute activity to a job activity. What ever job is using that output value, define an extra parameter, say Script1ReturnVal'. Right click on its expression box and you will find the activity variable activated. Choose .$CommandOutput...

O thats what you mean. You probably dont have the execute command activity connected to your user variable stage. There are no incoming links and hence its not activated. Connect a link from your execute activity stage to your user variable stage.

YOu need to right click there (in the User Variables tab) and add a row first. Define your variabe (give it a sensible name). Then follow my previous post.