DSXchange

Simply don't do anything that would mess up the order. Computers are dumb - they do exactly what you tell them to do. You'd be well advised to run the PXPivot stage in sequential mode for this one.

What are you trying to do when these errors occur? Just open Designer, or something else?

I'm surprised you'd be attempting to write the DS_STAGETYPES at all; this should be static since installation (unless it's nowadays automatically updated from XMETA).

Welcome aboard. I don't see any duplicate in your example.

Well that's not horribly difficult, but you can't use $PROJDEF. Getting the value from the environment can be done via the env command (perhaps called through DSExecute), then that value can be used as required. If the value is not available in the shell environment, then it could be read directly f...

$PROJDEF is only used as the default value of a job parameter.

To have it evaluate, simply don't provide any non-default value for that parameter when requesting a job run.

I should probably point out that you have design-time control over how much Balanced Optimization pushes into the source and target databases. For example, you can specify a stage in the design at which to stop Business Optimization from doing its thing.

Use a server Sequential File stage, either in a server job or in a server Shared Container in a parallel job. The server Sequential File stage can handle embedded newlines; the parallel Sequential File stage cannot.

Scroll right in the Columns grid to find the "contains newlines" property.

My solution is to use a prior server job to determine whether there are any rows to process. This job stores the count of rows to be processed in its user status area, which drives a custom trigger ($UserStatus > 0) to determine whether the next job runs or not.

You can (should) find this information in Connectivity Guide for DB2 Databases Table 17 page 88 (for version 8.7).

That's what I had in mind when I mentioned "grep and two cut commands".

Boolean = 1 is one of my pet peeves. You don't need the "= 1". DataStage expressions do support a Boolean context. Therefore, in this example, the stage variable (TinyInt) derivation can use input.keyChange alone (its value is only ever 1 or 0). If input.keyChange Then sv2 + 1 Else sv2 Si...

Teradata.

Pre-check that there are records to be processed; don't attempt to process if this count is zero.

Is it the case that there is at most one tag per physical line? If so, test the line for beginning with a tag (or even with a colon); if not, simply accumulate the string. Form groups based on the tag and preserve only the last line in each group, probably with a Remove Duplicates stage.

Check the man page for cut - it's a really useful command to have in your repertoire. Yes, awk or sed could also retrieve this value, but they take longer to learn. There is a solution with grep and twocut commands in the pipeline.

It should work as you say. You have exactly matched the spelling and casing of the routine name. Try compiling the routine again, try adding the routine to the job's Dependencies grid (in Job Properties). Make sure the routine is receiving the correct number of arguments. If all of these fail involv...