DSXchange

Almost certainly the problem is that a database deadlock occurred during the operation.

You need to get a DB2 DBA (assuming your XMETA instance is DB2) to trace where the deadlocks are occurring.

(From memory) isn't there a "treat warning as fatal" property or something similarly named? A "fatal" will issue a rollback in addition to aborting the job.

Of course. Accumulate your totals in your output derivations, then run the whole lot through an Aggregator stage grouping on Item ID and passing only the last row of each group.

I second what Vincent said. The fix pack contains six (?) bug fix versions of 8.0.1 all rolled up into a single installable image. Though I didn't experience any problems installing it (Red Hat Linux server).

You also need a business rule (CENTURYPIVOT) that defines how to handle two-digit years. If the volumes are not large, a BASIC Transformer stage can do this easily. But what format does your target need to be? Here is one example (for a BASIC Transformer stage) that generates an ISO 8601 format date...

There are at least ten different definitions of "Julian". What is yours?

What's different between the two environments? ("Nothing" is clearly not the correct answer.)

Please post the precise error messages that the job generates when run in the second environment.

Among other things the script sets appropriate permissions on executables in $DSHOME/bin. These executables' permissions must be correct for other users to be able to use them.

These scripts are regular shell scripts - you can view them to see what they do.

What happens when you execute MSScriptDir/MSRunOraProc.ksh outside of DataStage, but with the same arguments?

Please note that dsrecords is an independent command, not an orchadmin sub-command.

Waiting for some other resource, such as I/O, network, a semaphore or database lock, will not consume resources (and therefore will not contribute to prstat and similar reports). But it might still engender a timeout. For example, trying to start a section leader on another machine on a saturated LA...

The only thing you can do is to rewrite the laws of arithmetic/algebra. Please solve the following pair of simultaneous equations for any constant k. kx / y = 5000 k(x + 8000) / y = 100 Of course rows/sec will fall when the rows are very much larger!!! <rant> Rows/sec is an almost totally me...

Navigation in the hashed file is by field number, not by field name. In each of the jobs open the hashed file stage, get to the Columns grid and scroll right. You will see something called Position. These must be the same for each field in each hashed file stage, even if some are omitted. I suspect ...

How were the hashed files created? Can you list their file dictionaries; at least the I-types that contain the key columns? (You may need to create VOC pointers to use LIST DICT.)

Which only leaves me the option to answer as follows:

0.4