DSXchange

The first one the file pattern matches. And by 'first' I mean the first one a string sort brings up, i.e. the same order you would see with an "ls" using that pattern.

And now I see we're talking about 8.x, where there's no such limit if the logs are being stored in the XMETA database repository, just normal db tablespace issues.

They are stored in dynamic hashed files so have the same size limitations and error characteristics as a 'normal' Hashed File stage in a job would.

UltraEdit: Edit -> HEX Functions -> Hex Editor

Right, checkpoints are the key here for restarts. And of course you could always run an individual job directly from the Director client, if needed.

Or, as noted, the DSJobController macro can be used just like a job parameter. Got a chance to look it up.

What's actually used under the covers is sqlldr and yes, much like DataStage it can read the file faster if it is fixed-width. That, however, doesn't always guarantee that it will have any effect on the load speed. And you really shouldn't have to set $APT_ORACLE_PRESERVE_BLANKS to False as that is ...

The "CMD" shell that Ray mentioned *is* the "DOS" you remember. And yes, you edit the plain-text uvconfig file using the appropriate editor of your choice (like 'edit' from DOS or Notepad from the GUI) and then regen it using uvregen from the command line.

No worries. Could you do a couple of things, please? It would help future searchers if you posted how you were able to solve this and please mark your post as Resolved using the button at the top of the screen.

ramkandavelu wrote:In the ealier versions of datastage i have used xml reader stage and xml writer stages ..but cannot see any of these in V8.1 version

Just for the record, these were gone well before the 8.x release, replaced by the stages you see now.

A known bug with an available patch, yes, as noted in that same thread.

There should be a Macro for that - and it will have 'Controller' in its name, from what I remember. Can't check right now or I'd post the exact name.

Generally considered to be an Oracle bug.

You basically have three ways:

1. Use one of the date macros
2. Use a job parameter, feed in the date
3. Write to a fixed filename and rename after job

The answer can be driven by how enamored you are of that particular date format.