DSXchange

Hi Guys, I am currently designing an ETL process. One of the main challenges that the client is facing, is the dynamic nature of the input (Oracle). The process should be smart enough to detect a change in the metadata and ripple the effect through the jobs all the way towards the end. I know its no...

I think its by passing it. The failure is being handled in the second trigger. If you want unconditional, then explicitly provide them there.

You can use case statements in basic routines. No case statements in the transformer, but you can achieve the same using if then else.

No, column names is not considered a row when getting row counts through DSGetLinkInfo(). You need to check for row count zero.

well then instead of checking for 0 bytes, check for 2kb. And if a file is 2kb or less, then delete it.

You can go to the sequential file stage properties and uncheck "First line is column names" option. That will truely keep the file empty if its empty. Look at Rays routine. He is using the STATUS function to determine the file size. STATUS() function returns different information in a dynamic array....

Cannot be done using datastage, as far as i know. You need a third party tool to do that. Google it, you will have some useful finds.

Wow, thats an accurate reply. I have an IBM dude here at my client place. I heard from him that its due in the first quarter of next year, or somewhere close to that time. No one seems to have an accurate answer. O well, the time will come

Here, Ray created something like this a while back. Take a look
http://www.dsxchange.com/viewtopic.php?t=101158.

You are correct. There is no option. You will have to write you own routine to detect a file with size zero and issue a delete command at the os level.

Refer to thispost

Double click on the phantom error so that the rest of the error message is revealed. Post that here.

How is it possible that the input is a valid, non-null date field and is not showing up in the sequential file. You must be constraining it in your transformer stage.
Another wierd thing, how come its showing up in the database and not in the sequential file

A while back i wrote a unix script and a basic routine to do almost the same thing. The routine wont work for you but the script will help. Check out thispost

What is your source and target???
The date format seems to be ok.And if Oconv(Time(), MTS) is giving you all zeros, then there is something else going on. Try loading the output to a flat file and check the format.