DSXchange

I'm trying to build this XML format where I have multiple repeated elements nested under <Customer> for each unique Customer Number. <Customer> <Customer Number='123456'> <Name> <Name Location="First">Greg</Name> <Name Location="Middle">S</Name> <Name Location="Last">Br...

Got a bit further, turns out my XSD needed the "unbounded" clause at the top hierarchy level to expose it as a selectable element, then all elements underneath became available column mappings <xs:sequence> <xs:element name="Customer" maxOccurs='unbounded'> <xs:complexType>

I'm attempting a simple XML mapping to learn the new XML stage. I've selected "top" to get all incoming rows combined into one single XML output, but when I do that I don't get any column mappings available. If I select "More.." and try to manually choose the column mappings I ge...

Was this ever resolved? I'm now running into the same issue. I need "File Name" as one of the output columns to use in the file name generation, but I don't want it as a column in my output file.

I want to make sure I'm setting up a particular service runtime scenario properly. I have a web service that I only want running if there's a submitted request. Then, from that point leave the instance running for any immediate subsequent requests, but if one hasn't come through within 5min then shu...

Is it possible to use a variable label name in a [GOTO <label>] statement? I've tried different things and it won't compile, so I'm wondering if there's a BASIC trick to doing this. Basically, I want my label to be a job parameter that is assigned at runtime and the GOTO statement goes to that label...

Thanks for pointing me in the direction of the Aggregator! Still trying to transition from Server to Parallel and just discovered the RowCount() functionality in Parallel Aggregator which is now working with no warnings.

Thanks for your quick response!

The aggregator is basically just serving as a row count. There's one field for "GroupColumn" and another for "CountColumn" where both are hardcoded to one. Then in the aggregator stage it's grouped on the "GroupColumn" field then summing the "CountColumn" fiel...

Have a job with an ODBC Connector input, Netezza bulk load output, then all rows going through an aggregator stage into another ODBC Connector output. The job runs fine, but always logs this one warning "main_program: No field specified for reducer, statistics will be calculated for all fields....

We had a similar situation and it turned out that our distinguished names we set inside WAS security settings weren't the exact same upper/lower case as how it appeared in Active Directory. Once we fixed that, it worked! Review your settings again and be sure you're not also victim to case sensitivi...

Thanks for the tips! After toying with it quite a bit, I'm getting the best results using 2 ODBC Connector stages (DB A & DB B) into a Join stage then a transform to drop the records that already exist. It takes about 30-45min to get through all 18,000,000 records and I'll continue to learn and ...

I have a job with an ODBC stage that is submitting a user-defined INSERT INTO...SELECT FROM statement. Is there a way to capture the response from the database and write it to the DataStage job log? For example, "17,000 Row(s) affected"

You're right, I don't need all the unnessary updates and #2 approach is better. I've got a job design created using the Join Stage where I'm reading in database A and left joining to the key field from database B then have a IsNull() transform contraint. It's working, but takes approx 45min to run t...

New to parallel design...

I'm sourcing database A and targeting database B. How would I best design a job that only writes records from A that don't already exist in B? It's a fairly large dataset (18 million records) with one unique ID key field.

Unfortunately, I tried the XML importer and it setup the columns where my example would come through as 3 rows. /Individual/Name/@Location /Individual/Name/Name/text() produces 3 rows... First Bob Middle Alan Last Smith Wasn't sure if there was a sneakier way to code it so that I got one row Bob Ala...