DSXchange

Darn, i should really start reading the posts completely rather than skimming through it One reason why this might be happening is that if the table is huge (input in huge) and you have transaction size set to 0. Maybe the table temp size is filling up and hence slowing down the process. Try putting...

What is the array size and transaction size that you have specified in your target OCI stage?
That can make a considerable amount of difference in the way your job behaves.

But a word of caution it do cause the table to be in disintegerated form ( thats how I frame for my people here ) i.e. you may need to restore table inetigrity! What do you mean by that? I agree that if the load job is killed or terminated abruptly, it leaves the table in load pending state, but th...

Also consider doing a backup, project wide backup, before you attempt anything.

Look into the DB2/UDB Load stage.

What solution did you use pradkumar?
You can also mark this post as resolved now.

Its not able to execute sqlplus by the id you are using to run your datastage jobs. Check your directors log, the second entry, which shows all the environment variables being set. See if the oracle env. is being executed or not.

Version of Datastage was 7.5/ server edition.
Unix system was Sun OS 5.9 with 8 cpus at our disposal.
We were processing about 10M rows. This was just the historical load and then i had seperate jobs that used the DB2 API stage for inserts as they were just a few thousands.

Well, my case was just a general case. Lots of transformations, lookups etc etc. I did all that and then captured the results in a flat file. Then my final job would pick that file and load it to the UDB table using the load utility.

Confused. What is the delimiter of your flat file. Is it double pipes? How did you specify that in your sequential file stage in the first place?

Well, can u comment sth on the performance level of DB2 Bulk Laod? Performance is a relative term. Very relative, proportional to the size of data, engine power of your cpus(thinking in terms of cars ), number of cpus, network traffic etc etc. Waht is rate at which it can process rows? I have gotton...

I highly doubt you can update data using the load utility. As the name suggest, it can just load, in other words, LOAD, INSERT, REPLACE etc and not update. Updates will be handled seperately as a logged activity (DML).

Are you getting that amount of data on a daily basis???
If yes then its going to be update too right?
If this is for the historical load then i would say yes, go for the db2 load stage. Explore it, search here to set its properties correctly.

Yes. Increase the array size to about 400-500 and either specify transaction size to 0 (commit when job is successfull or rollback on abort) or you can specify a size of 10k. Depends upon you.

trobinson wrote:Sometimes I write stuff down so I guess pen and paper are also useful tools. A whiteboard can come in handy but is a pain to carry around.

Thats hilarious. What about a coffee mug. Does that count, hell, it helps keep me awake