DSXchange

Curiously, possibly not the case.

One of the tuning tips mentioned at IOD 2009 is to read with a Sequential File stage (sequentially) and parse with a Transformer or Column Import stage (in parallel) - this combination often runs faster than parsing in the Sequential File stage.

If it's done right none of the components gives any trouble.

There is not really any need for a WAS administrator; once WAS has been set up (primarily for user authentication) there's nothing more to do other than to include it in the proper shut down/start up sequence.

In server jobs error messages are always from active stage types (in this case Transformer), because only these have associated processes. When you reset (not re-compile) the job are there any "from previous run" messages logged? If so, what's in them?

Any column marked as Key in the Columns grid can (should, must) be used in the WHERE clause.

Does not the ODBC stage allow you to generate UPDATE statement?

Not true. Sequences ARE server jobs.

That's one of the reasons server jobs aren't going away any time soon.

In version 8 the better mechanism is to disconnect the session using Web Console for Information Server. When you open the job you get a dialog showing the owner of the lock. That's the session ID that you need to disconnect.

What database? Which stage type?

Select sorted in descending order. Constrain on input row number = 2. Read what I said about this not being an ETL task.

You could perhaps find out be examining the stage properties in the Stage Types branch of the Repository.

If not, the limit would be documented in the underlying table (DS_STAGETYPES) but whereabouts in that record structure is something that is not publicly documented.

Only implicitly. Within the Join stage you identify the "join keys" which would, in SQL, occur on either side of an "=" condition in a WHERE clause. But no other conditions are possible - it's a "pure" join.

Maybe, maybe an XML Input stage, maybe something else. I have no idea at all what you want to do with the "XML file" column. However, what you do have to do is to segregate it, and to do that a Transformer stage parsing a single VarChar field seems to be a good technique. Add a unique row ...

No way to know. Try the import again, and this time record any errors. Maybe restart the client machine first.

Well that explains why Field() can't find field #5 at least - it's not on the same line. You need a more comprehenive way to parse these data, perhaps do the parsing in a Transfomer stage that segregates types H, D, T and "other", with the others being your XML. Does the XML itself contain...

It's not an error. It's an alert. You are generating multiple output rows for some input rows - perhaps using a Lookup stage that returns multiple rows because the lookup is on a secondary key.

Not really without a lot of reconfiguration or tricky synchronization. Profiling takes place in a dedicated project called ANALYZERPROJECT by default, while a different project is used for ETL. In any case, the results of profiling are typically used to come up with improved ETL specifications. Once...