DSXchange

Server or parallel?

Check that all services do, in fact, restart. For example the dsrpc service may not start if there are still connections from before you stopped the services (you can check with netstat).

You can only have one output link from the Sequential File stage. What you are trying to draw is the second output link, which automatically becomes a reject link containing one column, of type Raw. The Lookup stage refuses to have anything to do with this, understandably. You need a separate Sequen...

Then, as noted earlier, any null field value that you supply must have the same width as the field would if it contained a non-null value. For example if the field is Char(4), then an acceptable null field value might be "####" or something of that kind (even four spaces). The null field value prope...

There are some interesting-looking "difference" tools in the Hawk release. The differences are hyperlinked back to, for example, the particular object such as a column on a link in a DataStage job design.

Only run it when nothing is open for editing, so that you don't hang up on locks in the Repository database.

You said it was a sequential file then you said it was a hashed file. Do you have a pad character set in the Columns metadata? If not, experiment with the pad character. Or run through a Transfomer to extract the leftmost character.

Do you have multiple telnet servers, each trying to bind port number 23?
For example, if you've installed MKS Toolkit, or CygWin, or any of the other UNIX emulators, they each have their own telnet server. Only one can start successfully.

If you right click on the stage there is a Grid Style choice on the menu.
De-select it.

Check first that you can compile the job sequence indepently.
When you can not, verify that the parameter reference you have used in a trigger expression is an accurate one.

MetaStage automatically captures these statistics from DataStage job runs and stores them in its database. You can report on them using canned queries or you can create your own. (ETLstats is cheaper, of course).

Clearing a hashed file is very fast - faster than deleting and re-creating. Under the covers it simply copies /dev/null over the top of the hashed file, then writes a new header indicating that the hashed file has 0 records.

"Year to date" might be a transformation, or it might be a summing of values for all dates in the current year up to today. Which are you trying to do?

Delimited or fixed width format?

Lots of places. It might even be a pathname in an ODBC data source definition. It's really difficult to say without knowing more about how your particular site is configured.