DSXchange

If you drop the connector entirely, and rebuild from scratch, problem disappears. Whatever old code is behind the scenes gets discarded when the new one is built.

Yes, after a complete rebuild from scratch, it's working. Exporting: that was one of the first things I did - could not find any reference to the bare T_Schema variable, except in the parameter section where $T_Schema was listed. As a result, I'm fearful that SQL code using old parameters hangs arou...

Well, I rebuilt the job from scratch, having only $T_Schema in the parameter list, and #$T_Schema# used in my SQL code within the DB2 connector. This is what the end result of the original job was (T_Schema gone, $T_Schema everywhere), yet the original job kept failing. Somewhere, the job/connector ...

Is it possible to update *certain* columns in a hashed file? I know writes to HFs are row-destructive (if that's the term) , but has anyone come up with an approach through transformer or whatever to update conditionally certain columns? In my case, I have a group of X records within a key, and vari...

deeplind07 wrote:Hash Partition (on F1 )and sort(and f1 and f3) in the input link of transformer. ...
SVF1_1=F1
IF SVF11=SVF12 THEN SVF2= SVF2+1 ELSE 1
SVF12=SVF11

Something similar to the above worked. Thanks!

Yes: It's feeding the 2 leading columns of a 3-column Primary key. Multiple rows from the target table may be deleted for each key that's fed, but that's OK. It appears there's an average 120 rows for each row passed. So in reality, the job is passing 4 rows/second to the DB2 stage, but probably 4-5...

When we do the subsequent insert, it goes around 200+ rows/second. When the delete occurs, it's feeding the 2 leading columns of a 3-column Primary key. Multiple rows from the target table may be deleted for each key that's fed, but that's OK. EG: 700K keys may result in 1.5m rows deleted. I will as...

I have a job that takes rows from a hashed file and deleted records from a DB2 (9.5) table. The target table has 122 million rows. The deletion process is very slow - rarely above 5 rows/second. Each week, the hashed file is refreshed, and contains about 1 million rows. The deletes are done in order...

Doh! That'll learn me not to paraphrase other people's text! Eric - I later tried: nohup $binDir/mkdbfile $projectDir/tmp_date_ids2 30 34183 4 20 50 80 1639 -32BIT 2>&1 < /dev/null & disown And, although I still got the 'disown not found' error, the hashfile seems to have been created! I the...

eph - tried: nohup $binDir/mkdbfile $projectDir/tmp_date_ids2 30 34183 4 20 50 80 1639 -32BIT 2>&1 & disown Got "ksh: disown: not found" (I'm on SunOS). What I'm seeing is - it just hangs - ie the process just exists (I check from another window), and it seems like there's no cpu a...

Hi, I have tried it within a shell, and also just the base command from the prompt. Same behaviour in a shell, or not. As soon as I tack on the & , it hangs. chulett - no, when run from command line, finishes on its own without further interaction. ray - I have tried with nohup. Eg: nohup mkdbfi...

I have a Unix(ksh) script that I use to build hash files using mkdbfile. If I run interactively, it works fine. If I run in background (using & ), it either hangs, or times out with a STTY-like error. Anyone know if anything special has to be done to get this command working successfully in back...