DSXchange

It depends entirely on the logic of the job. The "trigger" is just a column on the input link.....what you set it to is up to you.....a series of counters that (for example) lead to a column that is always "1" for the first 500,000 rows, and then "2" for the next 500,00...

Have only one link, going into only one XMLOutput Stage. Sort your data before going in. You should have two columns on that single link to XMLOutput. col1 /detail/officecode/text col2 /detail/officecode/dep Your xmloutput can be the terminating stage (for testing, this is fine...just put in a filen...

Then I am still not understanding the problem.

There are only two choices.

A source/target column in FastTrack either (a) matches the full 5 part name of already imported metadata shared tables or (b) it does not.

Ernie

Please explain better what you are looking for as your output...from what I can tell, you don't need the complex technique....you only need that when you have "multiple office codes" with unrelated "multiple department codes".......in your case, all the dep codes appear to "...

I don't know what else to tell you....look in the tree of database tables that you have in FastTrack.... You will see a 5 part name for each column.... host.database.schema.table.column If you fully qualify these names in your csv file being imported, then they will match....check spelling, upper/lo...

There are numerous ways.....one of the simplest is to import those target tables while inside of DataStage, using the "Connector Import Wizard" (as opposed to ODBC table definitions, or via older Plugin Stages).....another way is to use the ODBC3.0 metabroker in the Import/Export Manager.....

Ok...then be sure and import those target tables and their columns as Shared Tables prior to importing any .csv mappings into FastTrack.

Ernie

If the source/target names were in your original spreadsheet, check their spelling and case carefully......it should match, assuming that you already have those hosts and tables as formal Shared Tables.

Ernie

I haven't tried it in years, but in theory it should, "if" such a thing is still allowed by the MQ API........under the covers, the stage is just a "well behaved C++ MQ API application".......and the Queue Manager is a connection property. Try it and let us know what happens. Ernie

Yeah...well, it would have taken me 3 days to get it to work in C++ ; )

It wasn't that bad. Probably really slow, but an effective prototype or fine solution for relatively low volumes. DS/Basic does "text" very nicely. OCONV and ICONV make moving things from decimal to bit values very easy, along with Fold and such for simple breaking of strings into equal 3 ...

I like that idea.... couldn't you also do this without having to edit the Job? What happens if you (first export it to .dsx) delete the "debug" container, then import the "dummy copy container" of the same name and re-compile?

Ernie

I'll take a stab at it and post it if/when I get it working. JavaPack is probably a smarter way to go, or even a BuildOp, but this will be more fun.

Ernie

I'm not exactly familiar with the problem you are running into, so it may be more involved than just cleaning out the table, but I like to do deletions via FastTrack....it has the simplest GUI for it, but if you are running 8.1.2 (possibly this was all in 8.1 with FP1...I can't remember) you can do ...

Hi Craig.... hi-jacking this thread, or more like "re-opening" it to see if you or anyone ever came up with a solution for the "first part"..... has anyone written a UV/Basic routine to do just the basic level of base64 conversion? (the whole 3 byte thing...in one pass...no other...