DSXchange

As noted, there is no such thing as an "RTI job" per se, but a job can be "RTI enabled". So it's still all about the job design when it comes to performance - whatever that means in an ETL context. (had to throw that in before Ray did)

Sequential files mean just that - they are accessed sequentially from beginning to end. Something must read it to find the 10 records you want, be it the DataStage job or something beforehand like awk or perl. Another thought would be to bulk load it into a work table and then use sql to pull out ju...

Briefly, RTI bolts on to either PX or Server jobs, allowing you to expose either as a service or leverage services inside either. And yes, it does come with new stages in order to allow that.

It "should" but doesn't always work without it. try with whatever user/pwd you would use with the client.

Does it work if you pass in full credentials - server, username and password? Or just the server?

FYI - The online help in the stage is decent, there is the XMLPACK_20_Designer pdf in your Docs directory and lastly there's the XML Best Practices documentation from Kim Duke's website in the 'Tips' section. All are required reading for working with XML in DataStage, *especially* the last one. Well...

It would be great if you checked the documentation, the functionality is well explained there. It is literally as simple as I stated, though.

kranthi wrote:When I pasted the following sql in the before tab to drop constraints...

That's not "sql", as in not a sql statement - that's a procedure. Which is why the suggestion to simply call or execute it instead.

Sure, utilize the 'Trigger Column' functionality in the XML Output stage to do this. A column that increments per record would do the trick.

Not you, obviously.

Ray Daignault, I assume.

they have well defined special framework for handling exceptions ( Sending pager , Mail and sms to cell) Just as an FYI, this same 'framework' is pretty straight-forward to do in DataStage as well and - if you don't feel like rolling one yourself - you can download one for free from our own Ken Bla...

I'm curious how you can be expected to prepare any kind of comparison between tools that you aren't intimately familiar with? Or did you... 'volunteer'? I'd be surprised if you can find any site where both have been compared head-to-head already, but you never know. Put your google-fu to the test. C...

I realize what the posted messages say, just don't really make sense to me given the explanation of the job design. Seems odd that error would occur reading from an Oracle database but then I don't use ODBC, so perhaps that is a 'feature' of walking that path. [shrug] Curious how this all works out ...

Funny, but the first documentation I received (back in the day when you still got printed manuals) still said VMark on them.

That error should only occur when reading a delimited sequential file (hence the 'read_delimited' error), not writing to one. And that you have a metadata mismatch.