DSXchange

Do each of the checks in a separate stage variable. For each of these initialize a corresponding stage variable to contain the name of the column or columns that are the subject of the test. Use that on the output link that captures failed rows.

You can also use a Hashed File stage and include BY @ID in the Selection page. Note that data can not be stored in sorted order in a Hashed File. I don't know if this is possible (and can't check at the moment) but you may be able to use Type 25 as the file type - this is not a Hashed File but a B-t...

Do you in fact have only three columns (COL1, COL2 and COL3) on the link between the Transformer and the database stage?

Are you sure it's a "WAP server error", not a "WAS server error"?

IBM internal staff training is occurring now. Business Partner training is scheduled for the first quarter of 2011. Public training from IBM might make it late in that quarter (a "what's new" class may occur sooner).

Job design information is also stored in the project's local repository database, which some mistakenly call "UniVerse".

The original author is the one who marks threads as Resolved using the green button at the top of the screen.

Welcome aboard.

I don't know the answer to your question, but some of the manuals for version 8.5 can now be downloaded from the IBM website (start from the Information Center, or search for "Information Server 8.5") and you may find your answer there.

It's the job log we're talking about here, not a log file. I can't envisage any way that std::cerr could map to a hashed file or to a table in XMETA.

Probably not. Partitioning is about grouping - you have to group like keys so that each distinct value occurs on only one node. Assuming that your four Join stages have different join keys, it is therefore apposite that their inputs are partitioned differently. A Join stage - which requires sorted i...

Maybe it is, but compliments are much rarer. That said, there are two kinds of message handler; those that are local to a job and stored with the job, and those that are named and stored globally (and are available across projects). Each project may allocate one named message handler as the default ...

So do you want to get data from a web site - for example using the Java pack stages - or from a mail server?

You're both right; you both mentioned SurrID. Craig just neglected to mention the table name. It happens. It happens in DEV, then you fix it.

svDatePart:: Oconv(Iconv(Field(InLink.TheTS, " ", 1, 1), "DMY"), "D-YMD[4,2,2]") svTimePart:: Oconv(Iconv(Field(Field(InLink.TheTS, " ", 2, 1), ".", 1, 1) : Field(InLink.TheTS, " ", 3, 1), "MT", "MTS") svFractionalSecon...

No, no, a thousand times no! NULL is not the same as empty string (""). Anywhere!

I can work with "". I can determine its length. I can compare it validly against other strings. This is not the case if the value is unknown (which is what NULL is).