DSXchange

Is each routine of type "transform function"? Only transform functions are visible in the DS Routines list. I'd be a bit concerned, too, that you've overloaded the name of an intrinsic function (Num) in an Equate declaration. It's better, in my view, to use assignment statements or literal Equate de...

The order in which they *should* be run is entirely in the hands of the job sequence's designer. Job1 and Job2 are just names. They could as easily be called Abbott and Costello. You can't expect MetaStage to read the designer's mind. In terms of lineage, it's really a function of how well job desig...

Performance in ETL is principally measured as elapsed time.
In the job logs you will find start and finish times for the job and for its active stages.

Oracle logs are usually located in $ORACLE_HOME/logs. If not there, check out whether your Oracle configuration specifies an alternate location.

Why are you trying to do it from UNIX? Don't you have access to a DataStage client? The "approved" method to purge a log is using the Director menu: Job > Clear Log, then choose immediate purge. There's no UNIX command for doing it properly (that is, preserving the control entries). You can, from th...

Take a look at the SDK routines that manipulate keys, which create the SDKSequences hashed file if it doesn't exist. Mimic that code, or just use one of the routines once!

If you have an ODBC driver (compatible with ODBC 2.0 or 3.0) DataStage can use it. After all, isn't ODBC a "standard"?!

DataStage's main interaction with SAS is managed only in the Parallel Extender environment (DataStage Enterprise Edition).

Although it's not an answer to the problem you pose, you can resize the columns in the data browser by dragging the column boundaries in the header of the grid. Tip: Don't, even though you can, use double click if there's a large amount of data out there. It's got to process the entire result set to...

Routines only appear in the drop-down list if the Routine has been successfully compiled. You may also need to refresh the Transformer (from the menu, or use Ctrl-R). Equate is used in two ways, to define a constant or to define a single-line macro. The first two examples define contants. Equate PI ...

Use the pre-configured localuv data source.

User ID and password will be disabled, as you're already logged in to that data source.

This error code can mean that the connection has timed out based on the inactivity timeout setting. This is most often seen when you've left the client unattended for some time.
It might also mean that the DataStage RPC daemon (dsrpcd) is no longer running.

Provided that the sequential file has a "table" structure, you can access it using SQL (and therefore perform reference lookups) via an ODBC stage using the ODBC driver for text files.

Come back in about 2011 and tell us how you fared - this mechanism will be WAY slow!!!

Typically you're streaming data from source into the fact table (ideally via a staging area). What you need to do is to provide a mechanism - probably a dataset - containing the "reverse lookup" from the key values that are found in business data and the surrogate keys with which you keyed the dimen...

Design both the population of dimension tables, and the subsequent population of fact tables. Design, too, the business rules that determine which columns are significant in determining that a change has occurred and which are not (for example, you usually don't keep old telephone numbers). Leverage...

There's actually only one physical log, but you get multiple views (one per invocation) in Director. Purge the log and all the views magically vanish!