DSXchange

XML Output.

2) Your result is an artifact of using a small data volume. For large data volumes, you will get a quicker completion time using two nodes versus using one. I would say that one input record counts as a 'small data volume'. And what kind of parallel processing do you think would be going on in a jo...

Any chance this job tries to 'monitor' itself? Depending on exactly what you are doing, that could cause hangs or fatal errors.

It's just that they preserve white space so things like code indenting (and ASCII art) survive the normal 'squishing' process.

It's not a natively supported data type. Sometimes you can get away with casting to a varchar or longvarchar, or you may need to start using the XMLType methods built into Oracle to query it.

Some information for XMLType in 11g, check with your DBA what of that is applicable to 9i.

Well... that's a little misleading. They are still read, they're just not propagated downstream. Dropping them after input is not the same as not reading them at all, which is how I interpreted the OP's quest. Besides, you still have to fully specify them in the input data, something else the OP is ...

The question about the 'metadata repository' was asked in case you were asking about upgrading the internal XMETA metadata repository from 9i to 10g but seeing how that's not possible (only 10g is supported) it just confused things. So, out of curiosity are all external Oracle databases you currentl...

Can you post whatever errors/warnings you are seeing?

As noted, you cannot 'read selective columns' from sequential media, it's not in its nature. Also, I wouldn't call 40 columns 'many' or too many to make it any kind of difficult to get the metadata right. Sheesh. How long does it take to import the metadata for the file and then load it into the job...

Updates can be expensive but it shouldn't be a 'huge pain'. Are the 'key' columns you are updating against indexed?

srinivas.g wrote:When I correct the name with Rdate job is working fine. But I cant actually change like this since this programs are in production.

What is this 'correction' that you need to do?

From UNIX you can start a TCL session using 'dssh', then LOGTO your project and then issue that command. Start it either in the DSEngine directory or your target Project directory, for the latter you can skip the LOGTO. Then do the SETFILE. And note the VOC record must be the full path to the hashed...

I guess I should have said 'an actual Universe database other than the local repository'.

You don't need to set up a 'Universe connection' unless (I suppose) you are connecting to an actual Universe database. Here, the technique involves creating a regular hashed file and then accessing it with a UV stage - and all that UV stage needs to do that is a VOC record. Now, you've already got o...