DSXchange

What is the DDL of the table, particularly the key column?

Does your job partition the data streams according to the DB partitioning scheme and load appropriately?

This is a common error when RCP is used, as the lookup stage returns the lookup key column twice, once from the source and once from the reference. EDIT - Sorry, I was thinking that this was a response to another post where RCP and column duplication seemed to be the error. It doesn't necessarily ap...

That functionality seems to be gone at V8, I've noticed that in 2 separate installations.

What is your actual error - the heap size message is a warning, not a fatal error and might have nothing at all to do with your real problem.

I haven't seen that one before. What stage and what column is triggering this error?

Do you have RCP enabled? Also, in your PEEK stage, look at the columns and see if you have a duplicated name, most likely stemming from the lookup stage.

You will need to pass your job parameters in the form #ParameterName# to your external command.

The generated key in the stage is a BigInt and your key column is declared as a normal integer. I'd redefine the key as BigInt and, if necessary, explicitly convert to integer later in the job so as to avoid the warning.

The error message is quite clear, in your lookup input the column "name" is declared as an integer where it was previous defined as a string. Fix this problem and your dropped rows issue will also disappear.

Was "S9(S) COMP-3" a typographical error? Can you do a "view data" from your DataStage designer (I am assuming the HEAP error is a runtime one)? Also, could you cut-and-paste the complete actual error message?

You can use the command line from the ADMIN tool or straight from a UNIX shell (that would be "dssh") and execute a "LIST {filename}" with DS Syntax or use SQL syntax with a "SELECT" statement.

As of version 8 there is a Parallel Engine Message Reference PDF file which documents the messages; before that there was no such guide.

Do those 50 tables have the same columns and are they to be loaded to a single dataset, or do they have different column names and counts and get loaded to separate datasets? One option with DataStage is to write one job if all the tables have identical layouts and pass in the table name as a parame...

In UniVerse the table contents and the table description (DDL) are stored in separate places and are not necessarily the same. The DICTionary of a file can also be augmented by other items defined in the VOC file, as is done for such column names as F1, F2, etc. In this case the DICT for the file on...