DSXchange

In addition, recommended best practice is to use a dynamic configuration file with grid implementations, so that the grid management software can adapt if you lose a machine or two. There is, or is to be, an IBM white paper or RedBook on the subject. Ask your support vendor to inquire.

What are PREV"DSX.UTILITYABORTTOLOG" and PREP"DSX.UTILITYABORTTOLOG"_ABC and ABC"DSX.UTILITYABORTTOLOG" supposed to do?

These are not valid syntactically, and the middle one might explain barfing over PREP.

If you need concatenation, you MUST use the concatenation operator (either : or CAT).

If nothing else, post the text of any warning that appears in the job log. At least we get a shot at narrowing down the possible cause. Oracle has over 8000 different error codes, if I recall correctly.

Is it REALLY UniVerse?

If so, your assertion that ODBC is 'not compatible to UniVerse" is incorrect.

The dot between LIST and DICT is OK.
Were you able to generate this error twice?
There's no dot at the end of DS_JOBOBJECTS by the way.
If that doesn't work, try the following and let us know the outcome.

The file that it can not delete is '/u01/Ascential/DataStage/asclscratch/uat_edw.MAINT_USERS_SECR.bad' therefore its parent directory is '/u01/Ascential/DataStage/asclscratch' It may also be that the file was created by a different user ID, or is still in use. These are other potential reasons for b...

Simple economics. Supply and demand. Total demand for memory exceeds available supply. Either install more memory or reduce the demand. Reduce demand by creating more efficient jobs, running fewer jobs at the same time, tuning the memory demand for particular stages (such as Sort stage) or any combi...

You probably lack write permission to the parent directory on the production server.

Write permission to its parent directory is required to delete a file because the directory (a list of contents) has to be edited to remove the file entry.

DateStage does not have an internal format for timestamp. It does for date and for time, but not for timestamp. You can, of course, convert a timestamp to the number of seconds since a zero point, but converting the date portion to days and the time portion to seconds then performing the requisite a...

In SQL Server it's not necessary to use a function; the string literal will be correctly interpreted as a date, time or timestamp if it is in the correct (ISO 8601 standard) format.

gbusson wrote:When you generate a SK, unless you sort data after the generation, the last row has got the max value.

After you can merge data and sort the highest value.

The aim of the tail is to reduce volume.

That's an excellent performance tip!

The DB2/UDB Enterprise stage was written to conform exactly to the rules. If you have a Char(N) data type, it must be N characters long. The DB2 API stage is essentially the server DB2 stage rewritten to fit in to the parallel execution environment. It is less pedantic about Char(N) data types. Neve...

... and make very sure your hardware is up to it. Calculate the number of jobs a (N+1)*NP+1 where N is the number of stages and NP is the number of partitions. Now multiply that by 10. Can your hardware handle that many hard-working processes?

Compile the job sequence.

Look in its job properties, on the Job Control tab.

Magic! The BASIC has been written for you!

Can you please verify whether or not the job design is as simple as:


Can you try a similar job with a Filter stage rather than the Transformer stage?