DSXchange

How many values can there be, can it be more than 2? This is getting really ugly. Start thinking of an alternative route.

Well thats for you to decide. You can go with either max or min. Change the subquery accordingly. A part of it is as follows SELECT F.col1, (CASE F.col2 WHEN '*' THEN (SELECT MIN(sub_D.col2) from DataBaseTable sub_D where sub_D.col1 = F.col1 group by sub_D.col1) ELSE F.col2) ...

You said it provided delimiter as none ? Its a pipe delimited file so specifying a pipe as a delimiter would certainly help? As for the decimal, what is the length and scale of that field?

How are you passing the environmet variable inside the transformer, remember, no surrounding hashes are required for a parameter/env. variable inside the transformer.

It might be permissions. Try storing the result in some other file, instead of the replacing the original file.


The above command converts a.txt and stores it in b.txt

Even i remember reading here somewhere that 8i stage with Oracle 10g is not a good idea, they will start acting up at some point. Maybe thats your issue. Other jobs might still be working at the moment, but you never know about tomorrow. I suggest you give support a call, i bet they will hit this po...

Ok, that doesnt help much. Whats going wrong, the constraint deamon must be on a tea break. Need more info What is the constraint that you have specified? What is the datatype of your key column? You can have a simple in.Id>=100 AND in.ID<=150 for the reject link and the ! (NOT) of that for your mai...

This has been covered here before. You can do key increments at the database level as well as DataStage(via surrogate key generator or even stage variables.) Search this site for more details.

Reset the job and see if there are any messages that pop up from the 'previous run' that are helpful.

Need actual sample data with explanation of how each case is derived.

Ill throw some light on Ray's behaf, actually ill put you in the midst of light, here click on the smiley

Great. Thanks for sharing that with us. Now you can mark the post resolved. Others should make it a habbit to mark their posts resolved for ease of search.

How are you getting xx1 and yy1. How do you know you need to make 3 copies of each row. If that remains static, feed the input file the transformer with three output links, add two columns, duplicate the first three columns, add appropriate value for fourth and fifth column in the three links. Conca...

Well, it would have been if the comma delimited values were static. Then they could have been changed to columns and a pivot stage would do the trick. I dont know how the varying fields would be treated by DataStage. I thought about it, maybe not long enough. Did'nt want to burn any more brain cells...

Sorry, that was a typo, i meant to say the file was created, not the job. Dummy me