DSXchange

Also if you have single job with "upsert" option (update + insert) you may end up loosing processing time to check whether it is update or insert. Therefore I will sugesst you to go for two links instead of one. There is nothing wrong in going with separate insert and update but wouldnt th...

Aborting job even when you get any warning hmmmmm You need to be preety sure if you need this, as in many cases I have seen dataloss with warnings but its not always. So you may end up aborting your job in half way even when you get "informative" warnings. Do get rid of such a dataloss cas...

As mentioned in the additional info collumn you are working on server edition then as far as I know you won't be able to do that on a fly!

ray.wurlod wrote:It expects the format of an Orchestrate record schema. ...

Although just to point out, you may have cases of dataloss even if the job finishes with warning

dsscholar wrote:Thanks Craig..

Ray, I am preety sure expert like you must be having answer to this also

ray.wurlod wrote:

tagnihotri wrote:what is brain and whats the purpose!

This is actually one of the all-time favourite philosophical questions! Part of the bigger question of the mind/brain duality question.

Lavanya, If you are not able to get the suggested docs, let me know your mail id (you can send me message) will send them across to you! Dear tagnihotri(as mentioned in the reply msg) and other good hearted experts, Thank you very much for your help. In his msg, "Tag" has mentioned, "...

As I understand this, if you read the data from the dataset setting date fields as char the values are correct ? If yes then I will say there is no harm to right them as char only, guys please correct me if I am wrong! hi i looking at correct dataset only Raghavendra.. If i read it as a varchar or C...

One more thing, lookup stage by default (if you select auto) goes for entire.

ray.wurlod wrote:Technical point: on SMP systems, Entire shares the virtual Data Set, so that only one copy of it exists. This is an optimization strategy. Entire can be costly in MPP environments. ...

Have tried on a test job working fine for me !

Just to highlight we have two types of transformer in PX i.e. basic and PX. Also Search the forum there is ample floating around As far as i know , the PX Transformer stage implements a C/C++ like sub Routine. While compiling a job containing PX Transformer stage , the job compiler searches for the ...

Yes you can read same DS via multiple links! The reason behind SIG errors mostly is resource issues. Check for memory usuage and resource utilisation ... In running some PX jobs, we have had SIGKILL errors returned when we have tried to access the same input dataset at the same time - why is this? I...

To be honest, this sounds ghostly! as in if there is no processing from extraction to landing then how come -1 day :roll: Can not think of anything first hand, but if it wud be me I would read the date as varchar or char and try writting it to seq file, just to point where the issue is! hi tagnihotr...

I do agree with you, because from what I have seen the jobs abort with connection failure error! Connection failures would typically result in a simple abort of the job... don't recall any situations where it would result in a 'hang'. :? But then I've never heard of HiPortfolio or if there is anythi...

Also look in config file and check for the filesystem and mounts for the directories mentioned in for scratch and resource! Then we can talk Hi all, I have been assigned to tune some parallel jobs. I observed in some jobs reading from dataset is taking huge time. Pls advice me to increase no. of row...

Can you please elaborate on "But when the same records are viewed in Dataset stage the dates are existing in the previous day's date" :?: hi All.. I read records from DB2 using DB2/UDB stage and loaded into Dataset. In this some Date fields are there. When i view the records from DB2\UDB s...