DSXchange

You've got two choices with XML in DataStage out of the box: you can generate 'formatted output' or leave it unformatted. Formatted is what you've shown in your first example. It is more for people rather than processes as automated processes don't care about formatting aka 'pretty printing'. Person...

What forum would that be, exactly? There's no dedicated MVS forum here.

Technically, you won't be doing CDC you'll be doing CDD or Changed Data Detection. Search the forums for CRC32 and ignore any conversations where it is discussed for surrogate key usage.

In those other occurances should be conversations on using it for CDD.

kcbland wrote:Good, don't start.

I concur. But there *is* documentation on it for the morbidly curious.

You didn't find the ftp.pfd document in your 'Docs' directory?

Any platform, not just AIX or UNIX actually. And 'extension' as in 'minor update'.

The 'x' version denotes a special build of EE for Windows, the only version before 8.0 was release that allowed PX jobs to actually run on Windows.

subhashini wrote:Till now we didn't hv source data, but source system is not 10g , it is 9i.

Too bad, there's a loverly little package in 10g - the DBMS_LOB package - that makes working with those fields types much easier. Which is why I asked.

Another thought, something not quite as elegant or performant as Ken's but simpler to get to. Don't use the OCI stage for millions of rows, especially if all you are doing are pure inserts. Swap out your target stage for the Oracle Bulk Loader stage set to 'Automatic' mode. You'll need to arrange fo...

And I agree that there can be multiple places you'd need to check for an "idle timeout" setting. As noted, first would be your database as I find that to be fairly common. Some tools can get you around issues like that by a kind of 'keep alive' setting where they ping the database periodically to le...

sqlplus, alas no. However, now that 'TOAD' is actually 'Toad', yes! It used to be an acronym - Tool for Oracle Application Developers. Now it's just the amphibian, and you really need to spell out which version you have as there are several: Toad for Oracle Toad for DB2 Toad for MySQL Toad for SQL S...

Ok, now that that is out of the way... Do you (ideally) have an xsd for or at the very least a sample of this target file you are trying to create? You would then use that to import the XML metadata and that process would build the XPath Expressions for you automagically. I find those generally work...

Sure, it helps. You've got an Unconditional trigger between the two jobs which means you don't care if the first job fails. We mentioned this earlier, for the 'Automatically' option you typically just define the 'Success' path and let the Sequence decide what happens when there is a 'Failure'. If yo...

They should show in the log after the ORA-03113 if that was the case. Munish, are there any other errors in the log after the two you've shown?

Again, and regardless, your DBA should be encouraged to pitch in and help with this. They would be the ones to contact Oracle support, open a TAR, etc.

As noted, truncates are not committed and by the same token they cannot be rolled back. Once done, it's done and there ain't no going back. Now, if you are talking about other people issuing a 'commit' on their changes then yes, that will release any of those pesky locks that get in the way of thing...

Might be a little early to mark it as Resolved, unless your experiments actually solved your problem. If that's the case, please share your findings with the world so that others with the same problem can benefit from your blood, sweat and tears.

Thanks!