DSXchange

soumya5891, when you experiment by changing the sort order, make sure that look at the score, since DataStage will automatically insert a tsort operator at runtime should the links be sorted differently. You can disable this behaviour using the $APT_NO_SORT_INSERTION and setting to to "true&quo...

Hmmmm... does sound awfully "harsh" as an error message for insufficient access rights.

My first guess would be that the userid you are using doesn't have sufficient rights in the database to drop/build indices as the bulk loader is wont to do.

The commit frequency is set under "Transaction -> Record Count" in the connector stage.

Ahh, missed that from the original post.

No, there is no way for you to reset multiple jobs at the same time.

The "dsjob" command-line utility lets you do many things, including resetting a job. Putting this inside a loop in a script would let you do what you want.

No, not from the director.

I haven't looked into the database DDL scripts, but if you keep all of your databases in the same character set plus do the same for DataStage then no conversion will be necessary and correspondingly no warnings will be produced.

I would guess that your server and client are on the same machine and that the admin id has sufficient access rights to the DataStage directories while your user-id doesn't.

Did you follow Ray's suggestion. I haven't been to a site yet where this approach didn't work.

The call is the same on UNIX or Windows. The path to the executable is "{DataStage Installation}\IBM\InformationServer\Server\PXEngine\bin"

No, that command will truncate ALL files ending in ".ds" in the current attach point directory

You can also write a dummy job which uses a row-generator to generate one row, then a transform or filter to remove that one row and write to your dataset; that will do the same thing as the command-line would, it creates an empty new dataset.

One can remove a dataset from the command line with "orchadmin rm <descriptor>" and one can truncate the dataset with "orchadmin truncate <descriptor>"

- Is this a sporadic error or does it always occur? If it always occurs, does it always occur on the same record / after the same amount of data? - What has changed between the time it worked and now? Both in the job, the table, the database and the system - Are you writing large amounts of data? If...