DSXchange

Have you searched the forums? Vendor used to offer a "best practices" class (not currently scheduled). Most sites have such documentation, but it is not for contractors covered by non-disclosure agreements to redistribute.

No, it's a different syntax error! Looks like table name qualifiers need to be outside EVAL. Need to be trickier, and use aliases. select a.DESCRIPTION aDESC, b.DESCRIPTION bDESC from MyHashedFile a, MyHashedFile b where aDESC <> bDESC and EVAL "INDEX(bDESC, aDESC, 1)" <> 0;

The manual Administering UniVerse contains the best coverage.

So dump everything into two temporary tables, generating an extra column in the child table using Mod(rownumber,950) for each detail line for each header line.

In another job simply join these. The extra column gives you the criterion for creating a set of data to load to SAP.

That technique would work, but it's wasteful, and would require a priori knowledge of the maximum number of rows that might need to be generated from any one range (in your case 100 was sufficient - or was it ?!! How would you know (in all cases)?).

Which option from Import have you chosen (sequential file or XML table) ? As far as I can recall, you don't need to specify a delimiter using the XML table importer.

I don't believe that the Group field is used in the XML input stage. The Columns grid is a generic control; the Group field is used when the grid appears in a stage performing an SQL SELECT operation (to add a GROUP BY clause in the statement; non-grouped fields require set operations such as COUNT,...

ABCD_ on the source side.

Nothing on the target side.

Partitioning is preserved. Sorting may not be. Place a Sort stage between the Filter stage and the Join stage, with sort mode "Don't sort (previously sorted)". In any case, DataStage should have inserted a tsort operator if you did not specify sorting on the input link to the Join stage. Make very s...

Message handlers are not exported automatically; you need to make your own arrangements. Local message handlers (linked to a specific job) are in files called Local.msh which are stored in a subdirectory called RT_SCnnn in the project on the server. Named message handlers are in files called name.ms...

Welcome aboard. I assume "falt file" is a typo. Use Sequential File stage, XML Output stage, External Target stage. In the External Target stage execute your uuencode command and direct its output into a file, then run a script to FTP this whereever it needs to go. Optionally delete/move the data fi...

If you don't have a CRC routine, you can use one of the change detection stage types.

MATCHES only works in the RetrieVe query language, not in SQL. In DataStage/SQL (or UniVerse/SQL) only the LIKE operator is supported. And, yes, the double pipe operator is for concatenation. As far as I can tell, your syntax is legitimate. Have you tried it from TCL? WHERE fieldname LIKE '%' || 'va...

Use DSJ.ERRNONE as the second argument to DSAttachJob(), then handle the job status explicitly in your code.

Is this error occurring in one of your routines, in a job sequence, or simply when a job run is requested? If it's one of your routines, there is some path through the code where the variable named ParamValue is not being assigned a value. If you can't diagnose it post the code of the routine here (...