DSXchange

3 Steps for that line:
1. Display the Hex/Octal values of the characters to see what you have.
2. Determine which NLS character set you are mapping from and going to.
3. see if the character is legally mappable.

If yes, you have a problem. If no, then you have found your problem.

Yes, the "!" at the beginning of the line means that it is a comment.

"IF Status<1> = 0" means the first field of the Status variable, this comes back as one field only so is equivalent to "IF Status = 0" in this case.

Do not delete the files in that directory directly. Find the DataSet descriptor files and delete those files using the "orchadmin rm" command or do it from the DataSet Management tab in the clients.

Most likely that "/opt/IBM/InformationServer/Server/Datasets/" has filled.

Ahh, right. That too. But I still think the main problem is that space which is preventing the 2nd parameter to be used.

10 nodes on how many CPUs? If you monitor your system while the job is running which aspect uses the most - CPU, Disk I/O, Network I/O, Memory (paging or swapping)?

When you use "#p_FoldName##p_File #" as your file name, what happen? Normally this will work correctly. (If that extra space at the last hash mark is cut-and-pasted then that is the cause of your problem).

Bikan - you really should know better than to use "urgent" in your message and expect a response. Invariably the frequent posters will ignore posts marked "urgent" for a while. Ray will certainly add his comments regarding the true cost of urgent and immediate support. This is exacerbated by the fac...

That is because hashed files treat everything as a string, even if you declare the column to be "integer" it won't perform a conversion unless it has to. If you change your derivation to "In.MI_SUBMODULEOF+0" when going to the hashed file you will get an error. The solution is to ensure that only nu...

Where did you start TCL? If from UNIX and the $DSHOME directory, you need to enter the command "LOGTO {YourProjectName}"

Are you 100% certain that you don't have an "overwrite" mode somewhere?

For (2) use the input link name and stage name of the sequential file stage.

Welcome to DSXchange, TP.

Is this happening when calling the job from an execute job stage in a sequence job or from the Job Control section of a normal DataStage server job? What has changed in your system since Monday and are any other errors visible?

I have a very large dataset with just one VarBinary column ranging from 100 bytes to 65K bytes per line. This file contains every combination, including 0x00. I would like to split this DataSet into n-datasets with m-rows per dataset, i.e. I have 8,000,000 rows and want to split it into 80 DataSets ...

It sounds like the job is locked, perhaps by an aborted Designer session. Normally if you cannot see it in the Director window the locks should be gone. I always recommend starting the deadlock daemon, this background process will ensure that this type of problem takes at most 15 minutes to clear it...