DSXchange

Partitioning occurs when the upstream stage is sequential and the downstream stage executes in parallel mode. Re-partitioning occurs when the upstream stage is executing in parallel mode and the downstream stage is executing in parallel mode, but specifies a different partitioning algorithm from tha...

Is this an interview question?

As a general rule strings are never rejected.

But you have not provided any information about Null Field Value property or the data values in the file, so a complete answer is clearly impossible.

Welcome aboard.

There's not a whole lot available in 8.7. One approach that I have seen is the use of a before-job subroutine to lower the priority of a job (renice) if a parameter flag is set.

Decimal(2,2) is not possible. Precision must be at least one greater than scale. So, yes, request the DBA to make the change.

We prefer to use a strategy that uses "general terms" category (with a small number of sub-categories), and make use of a custom property to classify the term further. We're also contemplating using labels for this or a similar purpose. Both custom properties and labels can be used as sear...

Use a Constraint expression that tests how long it has been since the job started and send the row down an output link with a row count limit of 1. In a sequence remember the job start time using a User Variables activity and, at specific places in the sequence, compare this against current time. Ne...

Can you post at least the relevant parts of your FD ?

You can't have both.
If you want DataStage to compare specific columns, you have to name those columns (at least) and that, in turn, limits or rules out the use of RCP.

Increase the size of the Inter-Process Row Buffer so that it can contain at least one of your XML documents.

And even if there were you would need to be running Windows Server 2008R2 (assuming Information Server version 8.7 or later).

Also make sure that your Services tier machine (Linux) can "see" your client (Windows) machine; this means both resolving the machine name to an IP address as well as connecting.

In a grid the configuration file is usually a prototype, with placeholders such as $$Compute (a Compute node) and $$fastname (a node name) whose values are placed by negotiation with grid management software.

Even with generated SQL I place a simplified version of the SQL statement as an annotation on the design canvas itself (Courier new, colour royal blue, top, left, transparent, no border are my standards for these particular annoatations). Makes life so much easier when doing a walkthrough with DBAs.

Corrupted how?

Incidentally, the special tokens for default values of environment variable parameters have "$" prefixes: $PROJDEF, $ENV and $UNSET.

With a buffer more than one row can be passed between stages. Think of it as a producer-consumer model - the producer stage is writing rows into the buffer as fast as it can; the consumer stage is reading from the buffer as fast as IT can. The buffer mechanism can resist input if it becomes too full.