DSXchange

Search the forum to see whether anyone has had this problem in the past and whether there's a solution offered.

Check your PATH and LD_LIBRARY_PATH environment variables to make sure that the DataStage and Oracle libraries are findable.

Join stage requires the join keys to be identically named in both input Data Sets. Your alias is what's preventing the join from occurring - indeed should be preventing the Join stage from populating a drop-down list of candidate join key columns.

I'm a big fan of efficiency and ease of maintenance.

All that's required is to lose the milliseconds from a timestamp.

Iconv() and Oconv() don't work with timestamps, so that you'd need to do some string manipulation in any case.

You've tried to use mv as a variable name. The command must be a single string - you must pass it as such. Then, as Arnd noted, mv is not a Windows command; it is a UNIX command. Even if you're relying upon MKS Toolkit or some other UNIX emulator on Windows, you should use "SH" as the shell argument...

or

What are you supposed to do? Get the terminology right (I believe you mean Data Sets). Provide enough information to allow us to help you. Post the record schema for each Data Set, and the Join condition you used in the Join stage. Prove the number of rows in each Data Set using the Data Set Managem...

The problem is in YOUR ExecuteSQL routine. At line 77 there is a variable used to which no value had been assigned. That is, there is some path through the code where no assignment has been made to that variable, then at line 77 an attempt is made to use the variable as an rvalue (that is, something...

From the DataStage BASIC manual: delimiter evaluates to any character, including field mark, value mark, and subvalue marks. It delimits the start and end of the substring. If delimiter evaluates to more than one character, only the first character is used. I tried it with two, where only the second...

Please inspect your result file with a hex editor, or Oconv(String,"MX0C"), and show us that the Char(10) characters have gone. After all, it's not a printable character.

The Sort stage can generate two extra columns (key change and cluster key change). You need the first of these. Pass it into your Transformer stage and use it to reset your counter (a stage variable, probably) to 1 when the key value changes.

Have you tried LongVarChar?

KadetG wrote:Thenk you for replay. But I need use datastage for that task!

Why? Who said?

Are you sure?

It's my understanding that Field() only allows a single character as the delimiter.

But I'm always happy to learn.

I now have more than 1 lakh of meaningless points...