DSXchange

Who says you have to abort the job? Make them justify such an ill-advised decision! Robust ETL requires that you maintain control at all times.

Welcome aboard. :D Your problem is that you can't know in advance how many represents 10% unless you know how many are in the file to begin with. You could pre-process the row to get this count. Direct the Reject output link from the Sequential File stage through a Transformer stage which tests whet...

Welcome aboard. :D

That's not a question we can answer.

You have to know, or find out, how (that is, according to which standard) the data are encoded. That will indicate which NLS character map is appropriate to use.

UniData natively is, like UniVerse, a non-first-normal form database. UniData tables can be presented, for SQL access, as first normal form tables using dynamic normalization techniques. A UniData Administrator is usually involved in this process. Once there exists a schema, then table privileges mu...

Presumably the UniData file has been SQL-ized and you have not been granted SELECT privilege to it. Ask your UniData Administrator.

Hi ArndW, Many thanks for your reply. Thanks a lot. I was not able to understand your last comment "I think that the restriction that job parameter values cannot be changed from within the job came in at Version 7; so if you used this functionality in the past you will have to redesign those jobs t...

I think that the restriction that job parameter values cannot be changed from within the job came in at Version 7; so if you used this functionality in the past you will have to redesign those jobs that used it. Job parameter values have never been able to be changed within jobs. Arnd is confusing ...

Loading pauses while the structure changes. MINIMUM.MODULUS shifts the structure changes from population time to creation time, but the same amount of work still has to be done. Meanwhile the timer is still running. This helps to explain the low initial rows/sec value. There are no hash collisions a...

I believe there are a couple of views on the MetaStage directory that can be used to get metadata out more easily than trying to query the underlying tables (which are complexly interrelated). Not working at a site with MetaStage this week, so can't tell you the names - check with your support provi...

jshurak wrote:BUMP

I find that totally offensive. I asked you reasonable questions to help to diagnose your problem, you have not answered them, and you are demanding answers. Also, we are people, not pinball machines.

Start by defining "performance" and how you measure it. Rows/second is not a valid mechanism for comparing between jobs for a number of reasons.

Split the date into year, month and day. Subtract 2 from the year. If the month and day are 2 and 29 subtract 1 from the day. Re-assemble date.

SetNull() can be used provided the target column is defined as nullable.

SIGKILL isn't called from DataStage. This indicates that someone has used kill -9 to kill one or more of your processes.

Please don't ask for "urgent" or "ASAP". As volunteers we find it offensive. Urgent help can be obtained from your support provider in exchange for large amounts of money.