DSXchange

No, I cannot. I have no idea what analytical tools are available on your mainframe. It's not important anyway; discover heuristically how much processing capacity is needed. You'll probably be pleasantly surprised.

There is technically no such thing as "the 51st row in a table". One of the precepts of normalized databases is that storage order is under control of the engine and neither controllable nor predictable. Assuming therefore that you have a key by which the data are ordered during extraction, the task...

If the file name is a job parameter you can make this test in a simple expression if all you want to test for is "contains eight contiguous digits". If you want to check that they form a valid date then you will need a Routine.

DataStage EE MVS edition can deal with fixed width flat files, delimited flat files, complex (multi record format) flat files and DB2/390 tables. There is also an FTP stage that can be used to transfer flat files to a UNIX or Windows based server. No other edition can capture files from the mainfram...

I note that uvregen is trying to generate a shared memory segment 6,924,252 bytes in size. Usually this shared memory segment is under 1MB. Please check that this large size of shared memory segment is possible on your system. Most UNIX systems have the kernal parameter SHMMAX, but AIX has more self...

You only need one stage variable, let's call it svLOG. Initialize to 0. Derive as If InLink.CPD = 0 Then InLink.LOG Else svLOG Make sure that your source data are sorted properly (by ID then by CPD) and that CPD is 0 in the first row. Make sure also that all values in each combination of ID and CPD ...

Yes, on a cluster of four Linux machines is the way that many of the Ascential PX classes were taught. And, yes, it does parallel processing.

No news is probably good news. The longer they spend getting it right, the better the end product is likely to be. Stay tuned, be patient. You are too late to enrol in the beta testing program. Have your vendor monitor the situation. You can be certain that it will come out with much fanfare. My tip...

Most segmentation faults relate to trying to access a memory address that you don't own. One way you can do this is to try to stuff more than N characters into a string[N] or string[max=N] field. There are other ways.

I would have to add to the answer to point 2 that the parallel Transformer stage will always out-perform the BASIC Transformer stage achieving the same task, for a reasonable volume of data. This is because to use the BASIC Transformer stage it is necessary to convert the stream of data into the "ty...

There are few "repeat offenders". I have a standard response, with a copy to admin@dsxchange.com, after which some apologize for stepping over the line. On two occasions my professional services have actually been engaged.

On a number of occasions advice has been posted that approaches made via private email or private message may be regarded as requests to engage professional services. Could this advice be reiterated in the Contact frame of the profile viewer? I believe this is where most who undertake these approach...

Aha, but functions are implemented as subroutines. So you can! The following are equivalent in DataStage BASIC. SUBROUTINE Name(Arg1, Arg2, Arg3) * code here RETURN FUNCTION Name(Arg2, Arg3) * code here RETURN(Arg1) It is for this reason that a function may have at most 254 a...

To use MAXIMUM you need to craft your list of values as a dynamic array - that is, put a field mark or value mark between each. For example

The original question was "in terms of performance". There's a break-even point somewhere between small and medium data volumes where the startup time of a parallel job means that it takes longer to complete than the equivalent server job. For medium to large, and beyond, data volumes, a well-crafte...