DSXchange

ray.wurlod wrote:However, the OP stated that this time it's a CHECK constraint.

And? A 'not null' constraint is one type of check constraint.

rsrikant wrote:By the way is there anyway to copy a snapshot of my job into this post?

Next to last post on the first page of the topic.

That 'SYS_' prefix just means whomever modeled the database didn't give the constraint a name. But you're right, the most common constraint is a 'not null' constraint.

In the Schema Browser - select the table you are loading and it should be listed under the Constraints tab (on the right hand side) for that table. Either that or find it in the Constraints tab on the left hand side.

Perhaps there is a difference between Sleep and SLEEP? I've never tried to use it all upper-case.

Check with your DBA, or a tool like TOAD (if working with Oracle) can show you the constraint definition so you could see what you are doing wrong.

Typical example: a field might be constrained to only accept a 'Y' or an 'N'... and you tried to insert a 'P'.

Last one in wins, so...

Sounds like a sql problem to me, not a DataStage one. Isn't there someone on site that can help you write your source query?

Sorry... better start typing.

Or click on the ellipses (...) and click your way through it.

There's no magic bullet. You either need to do it in your source query (like with NVL for Oracle) or in the derivation of each column.

Check out the transforms NullToEmpty and NullToZero- they do that substitution and make your job a little easier to read.

By upgrading.

The substring operator will work from the 'end' of a string. YourField[1] Tells it to take the first character from the end of the string. You could then compare that to "-" and if equal, swap it around to the front. Hmm... you posted in the Server forum but marked your job type as Parallel. Are you...

meorz wrote:is there a book, online manual or anything about DataStage?

Book? No. Only 'official' things are the online help available from each Client tool and the .pdf documents automatically installed onto your pc with the client.

You can (and must) use a single ODBC stage. The only time you would need to use multiple stages is if the target tables are in different databases. Each link can point to a different table, the only 'restriction' when using a single stage is they all must reside in the same database / instance and b...

Use the same patch.