DSXchange

No. That only removes the control file, not the actual data files on each partition. You need to use the orchadmin utility (with its rm option), or use the Data Set Management utility within the Manager client.

Yes. Even though UV stage identifies columns by name, the underlying engine still must use field-number-based navigation. It's akin to using partial table information in other database stages; the columns you don't name aren't accessed. Therefore adding columns to the table will not affect existing ...

Don't re-invent the wheel. there are already routines supplied with DataStage for performing this task. They're part of the SDK, under the Key Management category of the Routines branch.

Yes and no. The term DataStage uses is "export", which creates a pure text image (DSX or XML format). When imported on the other server, this resolves into the appropriate structures within the DataStage Repository there. The point you've missed is that there is no single "deployment directory" crea...

Memory allocation for Aggregator stage can be tuned from the DS.TOOLS menu. It will fail if it runs out of memory. One way to avoid this is to sort the data by the grouping keys; in this way, the Aggregator can output the results, and free memory, whenever a grouping column value changes.

Did you ask the Ascential person who advised using OLEDB how to trap errors using it? If so, what response did you get?

There is a separate forum for QualityStage, you might benefit from looking there. QualityStage performs four tasks; investigation of your data to determine how comprehensively your rules cover the data, standardization of data into standard forms (for example "Av", "Ave" and "Avenue" can all be conv...

Stay tuned, we're actively investigating taking this on the road. (Still no free minibar though.) Australia is likely to be the first non-US presentation.

In parallel jobs provision is made for a final delimiter as well as a record terminator. This allows data terminated with a character, such as ";", to be managed properly without having to declare a dummy column and manage multiple field delimiters. There are, therefore, two properties. Your case, t...

0 rows/sec is not the same as 0 rows. These rates are rounded to the nearest whole number, which means that the link finished in less than 0.5 sec. Another reason I believe that rows/sec is not a useful metric.

That's basically because there is no mechanism for doing so. Essentially you will need to touch down a list of file names to disk, which the subsequent job can read. Of course, you could pass the XX parameter to each of the jobs run from the job sequence, and allow it to be used to construct the fil...

ashwin141 wrote:Hi

You can get ppts from Ascential(IBM) for Datastage EE also. Otherwise User Guide is the best option.


Regards
Ashwin

Can you please elaborate? What Powerpoint presentations are these?

Have you connected links to the stage?

The trick you need is to realize that the internal format of dates in server jobs is an integer. If you divide by 7 and take the remainder (the Mod() function), then the result is 0 for Sunday, 1 for Monday, through to 6 for Saturday. So you take the date (current date is available as system variabl...

Your job properties are open somewhere else, perhaps in Manager, perhaps in Director (Monitor view), perhaps by a different process. Did you shut down a client improperly? Whatever opened the job left a lock in the repository to indicate that the job was in use, and that lock remains. You will need ...