DSXchange

jhmckeever wrote:Thank God - I thought this thread was about to reveal some mysterious new "dsjob" behaviour I'd never encountered.

J.

I felt the same. Thats why immediately tried the command to double check at my end and results came negative.

Are you also doing OCONV along with your ICONV? I didnt know oracle could accept internal formats
Look into your log file, do you have warnings?

That looks good. Also try looding your excel to the lookup file set. See if you see any performance difference.

Thats what i was going to say, dsjob -ljobs just retrieves the jobnames without their instances.

Make it nullable in your target file and see what happens.

Search on the error message to find out what different instances can cause this to happen. Is this a straight pull from source to your flat file? Are you absolutely sure that this column is not null in source. Put a null check on that column. Something like if ISNULL(in.TYPE_ID) then "N...

There are so many functions that achieve similar tasks. I can use ICONV/OCONV and FMT function to format a string. Which one to use is simple the developers choice. You were getting problems with that function, I advised to use a macro that achieves the same functionality. If it works, well and good...

Build a lookup data set with those entries. Do a lookup. 2000 If then else statements are absurd.

Hmm. This way a person can increase the cpu's from 2 to 16 and keep running the jobs without fault, except for that one warning.

Search on keywords "prior component with the same name".

fmartinsferreira wrote:
Good, if some warning happen I will ask a new Authorization Code.

I cannot only have aborts. :D


Thanks

But I am not sure how much grace period you get with those warnings. Someone else can throw some more light on that.

Very smart analysis and suggestion by Ken. Two thumbs up, Ken

You can do that in TX

chulett wrote:Slave.

Maybe not.
Maybe a student.
On a serious note, I will definately let you know Kim once I am done with my current engagement. Thanks for looking out.
Regards,

If this were server then you would see a considerable amount of difference, but px sorting is pretty sleek. As Ray noted, sorting and agregations are still being done and hence your results.