DSXchange

I don't know of any variables that affect the DataStage code generator (the part that converts the transform stage into c++ code), after that the normal compiler options are read from the file and it might be something in the pre-processor which is creating different code. My guess is that the "...

There is something wrong with environment 1. Setting XXXX.CITY is missing when null. The inline if (ternary operator) is missing. If you scan through the code there should be a reference to XXXX.CITY somewhere else. I wonder if the difference might be attributed to c++ compiler settings, particularl...

Curious indeed. What is the exact derivation code in your transform stage for this? Are you certain that all is identical (particularly the nullability of the columns in question)?

ArunaDas_Maharana - that "20" is an optional parameter the SQL-Server "Convert()" which specifies the string format. It should be used to make sure that no incorrect parsing of the string date is made.

Custom buildops need to specifically define a schema for links. These are linked into the buildop via defining a "Table Definition", which has most likely not been copied over from the old installation. Open up the Build-Op and look at your tab Build -> Interfaces -> Input. There will be a...

The routine is either not catalogued (within the DataStage engine) or not catalogued correctly. This error message is most likely just the tip of the iceberg and it would be best to uninstall, delete everything from the filesystem (and registry) and re-install. But as has already been pointed out, t...

I think that part of the message meant that a call to free() (release memory) on a data type of "double" failed - indicating some sort of a pointer mixup since freeing up memory only fails if it tries to get rid of memory that it hasn't allocated. Can you start taking away single elements ...

I'd be surprised if that could be done. The problem is that all the stages pretty much get started at the same time (assuming a parallel job) so even if your small ODBC stage were to get the variable and write it to a file which is used as the SQL command for the other stage you get concurrency issu...

Is your SQL-Server datatype "datetime" or "datetime2"? Oh, I just saw the error - your convert() is backwards... plus you don't want to convert to a timestamp in SQL-Server, that is a very different type of field - you want datetime or datetime2 try convert('#ps_AppraisalFormSect...

I'm not at a DataStage session where I can confirm, but think that should work.

The reason is that DataStage doesn't need more than 20% of RAM for this specific job. The LOOKUP Stage has a memory limit to hold the reference data, but since your job didn't abort you didn't reach that limit. DataStage buffers rows between stages in memory, but that would be only for 100,000 rows;...

It has been 2 or more years since I did a DataStage with MQ project, but seem to recall having hit the same problem as you are seeing. What OS are you working on? We were on AIX and I believe we did something AIX-specific. There was also an interrelationship between the default and max block sizes i...

The first message is just a warning and isn't applicable to your problem. What is the value of #ps_AppraisalFormSection_DeltaDate.EXTRACT_TIMESTAMP# at runtime?

I think that the only way to go here is to either turn off purging until you get the latest version or use some other means to ensure that you don't run the sequences in question in parallel.

RAM, or "virtual memory" is per process. DataStage starts quite a few parallel processes depending upon the configuration file settings and how a given job is designed. Buffering of data is performed both in memory and in files, also governed by settings. Thus the answer to your question i...