DSXchange

You're not trying to set a parameter in the same job that's running?!!

Oh gosh, you are!!! (Notes use of DSJ.ME as job handle)

This is specifically prohibited. It's not possible. Job parameter values are set when the job starts, and cannot be changed thereafter.

Before we do that, please summarize what information you have actually gotten from the message. For example, I can discern that you get this message using View Data when still designing the job. I can discern much more. But we need to know how good your diagnostic skills are. Philosophy Corner If yo...

Please wrap your design in Code tags so that things align as you would wish. Use Preview to get it right.

Etiquette Note We don't do urgent here. It is an all volunteer site, we post as and when we can. And then only if we've something to contribute. If you need urgent, then sign up for premium service with your support provider, and learn the true cost of urgent. Or, as the old saying goes, lack of pl...

The space between the function name and its left parenthesis is probably the culprit here.

Error code -1 decodes as "invalid handle".

Fundamental Principle It is not possible to combine rows from multiple sources unless there is some common factor, without generating a Cartesian product. A Cartesian product contains M * N rows, where there are M rows in one source and N rows in the other source. You don't really want to go there.

No. Rebooting destroys the evidence.

Welcome aboard. :D

The Pivot stage is what you need. It has its own manual (located with all the other manuals on your client machine).

If the keys are integer sequences, then Modulus becomes a candidate for most efficient partitioning algorithm. If the two tables are in the same instance of the same database, a join performed in the database will almost certainly be more efficient than using any of the DataStage stages - the latter...

ambasta wrote:I cannot create any additional column.

You CAN create as many additional columns in your job as you require. Columns are link properties. These extra columns may exist only on the link between your Transformer stage and your Pivot stage.

Assuming there is a maximum number of values for the two-character codes, separate them into separate columns (probably in a Modify or Transformer stage using substring techniques), then use a Pivot stage to generate a row for each.

Welcome aboard. There is no Pivot stage in MetaStage, and no metadata implications for pivoting. Try posting a question in either the server forum or the parallel forum, depending upon your job type. You have indicated it's a server job, so that's where you ought to post. The Pivot stage has its own...

The safest way is to do two passes. One does the load, and captures any rejects. The second pass either performs a lookup against the target table to set "F" (all others in the staging table that are null are set to "L") or, if you have captured rejected rows, reads those keys to set the flag to "L"...

User-defined SQL could use the quoted identifier "Column A". Apart from that I can't think of any other strategy.

Put an annotation and/or some description text into your design to alert future maintainers of your reason for using user-defined SQL rather than generated SQL.

You have a special email address for reporting problems to the Hawk Beta team. They need to know. Let them know - the fewer errors there are in the GA (general availability) product, the happier we'll all be. I don't believe that Windows XP is a supported platform. But I may be wrong - I have not ye...