DSXchange

You are providing some value that is not suited for the data type into which you are loading it. The rest is up to you and your detective skills. You might consider using a reject link and capturing the failed rows.

Moving jobs, however, is straightforward. Export from one project, import into the other. The export file format is compatible into higher versions.

Can you please identify which stage (and its type, which I'm guessing is Transformer) is throwing this error, and what that stage is doing?

Are you certain that it's the Oracle stage that's throwing the error?

After all, you don't seem to be certain even of the stage type.

I am aware of a colleague who used a 16GB thumb drive for extra scratch disk when no other disk was available, but have no information about performance other than that it was successful.

OK, I'll take another look. Multi-select is OK for "mark as deleted". Will report back on the others.

Do the terminal $ characters have any function, such as "continuation character"? Does your metadata take these into account?

I exported BG in XML format from an 8.5 system then tried to import it into the 8.7 BG with workflow enabled. There were about 350 terms and several categories. They all went into the Development Glossary, as expected. There is no facility for multi-selecting terms and/or categories to take them thr...

Could it be that there's a Ctrl-M at the end of the lines (shows as ^M in vi) that is interfering with the parsing? That is, in what mode did you perform the FTP and does your FTP tool handle translating line terminators?

Was this file sourced from a Windows system?

"0x1a" is 26 in decimal, or Ctrl-Z. Interestingly this is the "end of file" character in (Windows) DOS files. It can be represented in DataStage as Char(26). Initialize a stage variable, data type Char(1), to Char(26) as its value, then convert to "" in your data: Conve...

You don't necessarily have to add checkpoints but, if you don't, you have to design your own logic to determine whether each job failed on its previous run and make the decision to bypass it or to reset and run it. This would better handle the case where the sequence did not stop (was not programmed...

I did not ask you to try to view the data in the Sequential File. What do you get in the Peek entries in the log?

Are there any warnings in the job log?

Without re-importing the table definition it is not possible. And even then the new metadata can not be propagated automatically in server jobs.

Short answer: it's not possible.

1. Define the rules that make them invalid.
2. Create logic to identify records that fail one or more rules.
3. Filter these rows.