DSXchange

Although it is possible to track down the approximate source line where this is occuring, it is usually much faster to look at your job, and in every transform look at the columns where you perform data manipulation, especially mathematical operations. As Sainath stated, most likely you are performi...

gpbarsky,

I think this name cannot be changed, it is the remote-function-call to/from SAP and is, I think, hard-coded.

Logic, Cobol fields are stored as clear-text except when they have an explicit "COMP" attached to the picture. The S9(6)V9(4) SIGN IS LEADING produces data such as '-3625418671' - a leading sign either - or space, plus 9 digits with the decimal point being assumed to be between positions 6 and 7. Th...

Since the Transform is the slow stage you want to avoid using it. It looks like you can't avoid using it to do your mapping of continents, so stick with the first solution that only uses one transform. I took a look at the column generator functionality to see if one could put a strimg algorithm in ...

jsgoncalves,

I would write a Sequence. use the ExecuteCommand stage to get the "dir...." results of looking into a directory, and then with the Loop (start&end) go through each element in this list and use a JobActivity stage in the middle to call up your job.

dsedi,

you are welcome - but could you post what the problem really turned out to be and how you solved it? That will make the search functionality work much better for anyone with the same or a similar problem.

Martin, one of the strengths of DS PX is exactly that which you are referring to, it is possible to write an ETL job which does not predefine the number or contents of columns. I think that you haven't gotten an answer yet was due to the way the comment was phrased; I understood it as either a quest...

Would it be possible to rephrase your question?

A DataStage routine can loop through each file in a given directory, but I am not sure what you then intend to do with the files found.

Ashwin, some very good points in your message. My gut feeling (inherited from performance number on Server) is that a sequential stage will be faster than a complex stage in this case where no complex file stage options are needed. The sequential file stage has to handle a lot less complexity and, a...

The problem is with the SQL privileges that you have (or don't have) to write to the UV_USERS table - most likely you will need admin rights to do this. What are you doing to trigger this error - I don't know where this might occur in normal use.

Mandar, I think this forum is a great place to have people exchange information and, as in this case, get some assistance in coding. I feel quite strongly that this is not the place to "give" out code. The forum shouldn't replace training but supplement it. It should be obvious that the statement St...

Kcshankar,

good news - at least now you have an explanation for the behaviour.

-Arnd

anu, Yes, you can do it without a function call. It is ugly and unmaintainable and not necessarily faster. The base logic is as Ray and I demonstrated, but instead of a loop you would use the STR({repeatingString},500) to generate your 500 rows per row input. I think that with your experience, the s...

kcshankar, I think that you have not really understood what both Ray and I have been trying to tell you. You need to perform 2 actions and checks: 1. Go to the DIRECTOR and look at the last log file for the last run of the job that only uses one node. Double click on the first entry which will tell ...

In the BASIC Programmer's Guide pages 6-248 onwards it describes the mask formats - this one denote 1 alphabetic character and then any number of any characters. The reason you cannot just do a COUNT of the file is that it contains several records that do not hold descriptions of jobs but control in...