DSXchange

Of course the DataStage hashed file caching works quite well - but I'm sure you know that and have reasons for not being able to use this. Depending upon how important this is to you, you can enable memory file caching for DataStage at the system level and use that; this would mean that your Hashed ...

I know that feeling - we have a job with over 250 stages with a strange bug but I cannot reproduce with a simpler example.... I'd rather spend my time making the end-user happy instead of making support/engineering happy with a test case.

Normally one doesn't see NULLs in COBOL, but they do use "HIGH-VALUE" and "LOW-VALUE" as indicators. The column in your original post that contains 0x00 0x00 is it a PIC XX?

But I am talking about restoring one job that uses a shared container. You can't just overwrite the hashed files.

Ok, so you expect someone to pick a column at random?

What database are you writing to? What has changed in the system since the point in time where this worked?

That COBOL copybook doesn't contain definition for the column "NO_PAKS"

iDomz - right you are with some LINUX packages. On AIX the truss command has been around since 5.1, and I know I've used truss on LINUX but it might have been aliased to another command.

That means that even if not using DataStage you have a mapping problem. Are you certain you are seeing Chinese characters and not just another Japanese character set? What are your instance and session NLS settings in Oracle?

If you are absolutely, positively certain that you are using the same APT_CONFIG file then it would seem that you are not doing anything wrong. Perhaps you are using 2 (internal) hashing-algorithms that do not use the same start point. Your first job hashes when going into a filter stage. What stage...

It probably is the reads. I've sometimes included my own cache in programs of this type by creating a Key and a Data dynamic variable for each file. The keys & data are appended as new values to these strings, and I check up on the Key (using LOCATE) before trying to do an actual read. In many c...

Both Linux and Solaris (as well as HP and AIX) have "truss". It may have been disabled or removed in your installation, though. I find it odd the a process going through uvsh is using all that CPU. Are the bad rows logged in a file or table or do they get written to the job log? Does your job have b...

Ag_ram, Because the dictionary name does not correspond to the VOC entry name, the VOC entry is not deleted, instead the data file pathname (field #2) is set to nothing, as per the message Can you please verify that this is the reason for the statement that VOC entry is not deleted? Why should Ray, ...

That is a strange code location - are you using encrypted parameters and/or $ENV settings in the job? Also, if you repeat the VLIST and grep out lines with reference to "$251" we can see where that variable is assigned.

I haven't seen the UNLOCK command wait on anything, it either releases the locks or fails to do so. Could you Cntrl-C out of the UNLOCK command? Perhaps you are having some sort of access issues; are you logged in as a dsadmin user with local Windows administrator privileges?