DSXchange

creating a job sequence to run 5 files usinh the same job to load into single table


used 5 job activity stages in job sequence . gave different filename in the each job activity.gave different invocation ids to each job activity.

the job activity stage still needs soem information . it still has the exclamation thing on the stage.

i tried compiling and it gave the below error

Expression : "Filename"- variable file1.txt is not defined.

Am i missing something to do?

please throw some light upon

thanks

Have you enclosed the filename within ""?
This may be the problem.

Minhajuddin ,

thanks it worked .

i am testing the logic using only with 2 source files .

the first file is loading very fast but the second job is really slow.

infact the second job is not loading into oracle but waiting to load.

After the first file loaded the second file is loading successfully .


should'nt both files load oracle table at a time?

it is taking the same load time as running the two files in a sequence.

I wanted to reduce load time using multiple instances.

Multiple Instances not necessarily equal faster or that they are able to break the laws of Database Loads. The first load is locking the table until complete, at which time the second can run. Multiple simultaneous loads generally require a partitioned table with each load stream handling a unique partition.

Me, I'd just concatenate all the files together and perform a single load.

I wish i could do that but the files we have are more than 30 million each .


so if i have a partitioned table , do i have to pass any paramters for parition or when i kick off the multiple instance job , it basically goes an picks a parition for each process ( for each file load).

Am i understanding it right?????

Do you have a partitioned table?

If you do, then it would be a matter of ensuring your loads were partitioned in the same manner as the target table, on the same 'keys' or fields. Unless you've made special arrangements, I'd be surprised if your files are arriving that way, so you need to take the time and space required to rearrange the data pre-load.

Otherwise, get your individual loads as efficient as possible (are you bulk loading?) then run them serially.

Tables are not partitioned.

i did run the loads using BUlk load stage.showed good performance.

But i was trying to do multiple instances running files parallely using OCI stage.

.