slow performance with Oracle CLOB field

Post questions here relative to DataStage Enterprise/PX Edition for such areas as Parallel job design, Parallel datasets, BuildOps, Wrappers, etc.

Moderators: chulett, rschirm, roy

Post Reply
AnushaReddy
Participant
Posts: 13
Joined: Thu Dec 02, 2010 2:55 am

slow performance with Oracle CLOB field

Post by AnushaReddy »

Hi,

I have a EBCDIC file and in that file i have different record indicators.
The last fields in file contains sub-columns ie., the meta data of the last field varies based on the record indicators.

Designed a job

Sequential file (with last column defined as VarBinary) --> Transformer ---> Load data to the Oracle table (the last field is CLOB data type in Oracle and defined as Varbinary in Datastage)
Its taking hours together to execute the job.
Array Size is sett to "1" in oracle connector.

Can you anyone help to improve performance of the job.
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

How many hours to process how many records? Unfortunately, pushing a CLOB through the job brings a lot of restrictions along with it, one of which is needing to keep the Array Size at 1... which means it will be "slow".

You may be better off converting the file into something that is "load ready" and then using sqlldr for the actual load step into the database.
-craig

"You can never have too many knives" -- Logan Nine Fingers
Post Reply