DSXchange

Running the orchestrate scripts directly from the command line rather than using dsjob or the DataStage Director might be faster, but only if the job generates tens of thousands of log entries. If that is the case it would be far better to find out why those entries are being generated and to get ri...

Ha - of course; once I read Ray's post I remember doing that sort of a multi-job compile many times in the past! I'd retract my incorrect answer if I could.

"DAYS SINCE" sounds to me like an integer column. The length of the output column is only relevant if you are writing to a sequential file.

Yes, a coldstart is preferable to shutting down and re-starting DataStage. I vaguely recall AIX filesystem issues with processes holding file-pointers to files which have been deleted and the spaces is shown as being in use and the UNIX file tables won't allow re-creation of that file but that was o...

Are you sure that the directory has been cleared at runtime? Are you executing a "CREATE.FILE" command manually or in a job? It might be memory/file system issue and before going to the lengths of re-creating the project a simple coldstart should be attempted. If you execute a command-line...

With Server job you can easily create you own maps, in Px that is not so trivial. Since the ISO-646 characters from 64 through 127 are invariant, they will be identical to their Latin-1 representations in 8859-1. The first 31 characters have no display equivalent, thus won't need mapping. The digits...

It would seem that the file "F:\DEV\REFPERS\PERS\HASH\HF_MDMPMERRORLOG" already exists when you execute your code.

I ws going to suggest redefine & pivot but this might be simpler and more efficient: 1. Redefine this as one field "Data" with length Char(1240) 2. get the MAX of INDEX(In.Data,'AB',1), INDEX(In.Data,'CD',1) and INDEX(In.Data,'EF',1) into stage variable "svFoundPosition" 3. I...

Yes, you should not use notepad or wordpad for analysis of this type.

I don't know of a way to do a compile of job-and-all-called-components (similar to the export options). It can be done with a script or a program, but the effort involved would be rather high.

Can the user-id you are having problems with login into DataStage and run the job from the director?

IBM has written an operator which takes care of ensuring that transactions are performed as a transaction; the operator was written by an IBMer named Julius Lerm for IBM Germany and I am sure that they would be willing to offer this solution, but for a price.

REPERROR = Repository Error

Do you have a DataStage admin who sets up the DataStage users?

Since the source has no delimiters, you will need to specify columns of fixed width. While the conversions from fixed to variable length are implicit, the one from "yyyy-mm-dd" to an integer is not, so your most efficient solution will be to use a transform stage between your source and ta...

This is a case for your support provider, as the functionality does seem to have changed and you have a good test case/example for the problem. Unless support comes up with a solution, you will probably have to use a conversion to string or similar in any case.