DSXchange

It's too late for backups, hope you have something to fall back on. Have you tried a "reindex" of the project? Search here for the details on usage and caveats.

Or something like that.

So... one line versus multiple lines?

Time to mark this Resolved it seem, using the big button at the top of the screen.

First off, it is a hashed file, not hash file or hashfile. Secondly, I don't see how one could ever 'rearrange data randomly'. Something else must be going on. Can you provide more details as to the design of your hashed file and exactly how you populate it? For example, is this field a Key field or...

Well, something is obviously different there. You need to keep digging and find it. Oracle client version, perhaps?

The format won't make one iota of difference, I'm afraid. The first thing it does anyway is convert your xml into dsx format 'under the covers' so you still end up doing a dsx import.

One reason why you shouldn't do large exports in xml format, your import time will increase exponentially.

Welcome aboard. :D I'm afraid it's one or the other as sql files cannot have job parameters in them. We 'worked around' this by having multiple sql files and using a combination of parameters to pick the right file at runtime. We only had a small combination of parameter values we would have used, s...

Exactly... or whatever your shell of choice would be. For Windows, the DOS 'start' command works from what I hear.

Then clearly it's not the same situation. Please start your own post with the details of your environment and the issue(s) you are seeing.

What database are you attempting to import metadata from and what drivers are you using for that?

It's normal behaviour as that's the place holder for the sign. What kind of help are you looking for? I could guess but I'd rather you be more specific.

Use dscmdexport for the whole project, dsexport for a single job but you'll need to script the latter to do singletons for the entire project. That or export everything and use a 'cutter' to hack it up, there's a perl one posted here IIRC.

There is no 'feature' within DataStage to do this directly or natively. What it can do is execute any O/S command, either from within the flow of a Sequence job or 'before/after' any job type. So that means any operating system command(s) or script that you have to do this could be run and monitored...

However, the 'File' path declared in the Data Set stage with the .ds suffix is just for the control file - not the actual Data Set itself. The actual data files go across your 'resource' disks from what little I know. Yes?

Post an example or two of your incoming 'timestamp' values.