DSXchange

I happened to read this while searching and it struck me that the suggestions are way off the mark. I'm sure you must have solved this by now but thought I would attempt to clarify the problem when others see it. So here goes; When XML is being parsed by the XMLInput stage, it must be read in it's e...

Can your Teradata installation accept (16 x 15) sessions? Maximum sessions is number of TPAs multiplied by number of AMPs per TPA. It certainly appears able to. I see rows returned for all 16 nodes with minimal skew in the same amount of time per node. I am marking this resolved! The problem was wi...

Are the indexes properly setup? BTW how much speed you want to attain with it? for me its a pretty good speed There is no where clause, it's a dump of a single table. I will verify that the table doesn't have issues by itself. 726 rows/sec from anywhere going straight to the bit bucket is acceptabl...

Nervezzo - I don't see one answer set from FastExport. I see many processes, all delivering data. In my case, I see 16 Teradata Enterprise Stage UNIX processes, all with 15 sessions per player. This means 240 sessions on the Teradata side returning into 16 nodes. It is my belief that Teradata IS ret...

Good idea.
bteq from the command line is equally slow. So from this I think we've eliminated DataStage. Must be network or teradata itself.

Could it be anything else?

I have a job that reads 1.4 million rows into 81 columns, 1000 bytes per row from teradata. It is a simple SELECT from a single table. The db is on a separate machine within a 100 ft of the DS machine. The db has 248 amps. The job reads with 16 players, 15 sessions per player. The output is to 16 Se...

"I'm intrigued by your views and would like to subscribe to your newsletter."

Have you taken a look at the Operators reference pdf for the changecapture Operator?

I assume your design is in the classic Kimball model and the compound key to the fact table is, in fact (hah!), the foreign keys to the dimensions. Therefore joining fact tables is really joining by dimension across fact tables - A federated Data Warehouse! Naturally all facts are at the same level ...

+150 Projects? Holy Schema file Batman! That's a lotta Projects. I'd go with sed too, if I had that many.

Why would you EVER touch the DSParams file Programmatically? Is there some problem with $ENV on a Windows platform? I thought $ENV allows one to capture the shell value of a user defined environment variable at runtime. Is this not correct? Assuming dsjob is wrapped could one select the db2 CURRENT_...

A join stage will never be able to combine 3 fields into a single row into a single field. The best you can do is pass all three fields in a output row to a transformer and build logic to choose which 4 char source field to put into a single 4 char target field. Going back to the SQL example, the on...

Server solution: the DBMSCODE will be available in the transformer after the link that goes to the Oracle OCI stage. This code value can be used in a constraint on a "reject" link to for example a Sequential file. If the DBMSCODE is non-zero, the row that just failed on the above link woul...

I don't know either but to find out I would;
Export a small Sequence with a Job activity with the option run.
Change the Sequence to reset then run and re-export.
diff the two.
The difference (the change from run to reset then run) "should be" fairly easy to discover.

I don't have Oracle available to me right now or I would test this myself and not waste your time should this suggestion be whacked but I gotta believe some combination of triple tick marks (definitely not double quotes) would work. '''A690','R456','F1234''' select * from table where codes IN (#p_in...