DSXchange

If you had specified tab when importing the table definition for the text file, all you need to do is click Load... on the Format tab in your Sequential File stage. The metadata are then loaded from the Repository. Tab will appear as a little black rectangle, but will work.

I'm sure I answered that a while back. Was it here?

DOS, funnily enough.

The switch for immediate return after completing the command is /C
Example:


SH works if you are on version 7.5x2 with parallel jobs licensed. This takes you to a "UNIX" shell - actually MKS Toolkit.

DataStage really only performs equi-join. Do the lookup based on FIELD1 and FIELD2 and make sure FIELD3 is returned. Perform the comparison A.FIELD3 > B.FIELD3 in a constraint expression.

Your support provider can supply these, and should do so at no charge.

They are also handed out (on CD) when you attend an IBM training class in DataStage Enterprise Edition.

It's in the Docs folder with all your other manuals.
C:\Program Files\Ascential\DataStage*\Docs
It's actually called Parallel Job Advanced Developer's Guide but it should be first in a name-ordered list as advpx.pdf is its file name.

Oops!

My answer was for parallel Transformer stage. In server jobs the correct answer is to use the Fix() function.

In a Transformer stage you can use the DecimalToDecimal() function. See Appendix B in Parallel Job Developer's Guide for full information. In a Modify stage you can use the decimal_from_decimal() function. If that's all you have to do, prefer the Modify stage, as it's much more efficient than the Tr...

Abdulwahabza wrote: I have a column in table whose data type is number 9(2) and for this field i am getting value from source as 12345.7626

Seems to me that your problem is right there! How can you get this value if the metadata specifies two decimal places? Perchance you have a metadata mismatch?

Yes it is. You need an additional output column for the count, set it up with Non Missing Values Count (or Missing Values Count) as its calculation function. If you need both, derive both and sum them in a following Transformer stage.

Right at the bottom of your code add the following lines.

"Phantom error" means nothing more than that the error/warning was generated by a phantom (= background) process. You need to reset your job to see whether any additional diagnostic information is logged.

Congratulations! It is often worth challenging from a position of relevant knowledge.

A "not equal" lookup would return most, if not all, of those three million rows for each key processed. Is that really what you want?

... but the count will be given in the status bar.