to change commit level

[kannan_ayyanar]

hi..,

I am new to datastage... please help me in this issue..

I am using ODBC enterprise stage in job. job is very simple, direct mapping from source to target...

Source is having 10,000 rows, if the job gets aborted after 5950 rows. the rows are getting committed in the target.

my need is

it should not commit in the target, so if i run the job again it should start from first row.

or

if i run the job again it should start from 5951 th row.

is there any option to set in ODBC stage or any other stage will help my need full...

[chulett]

You'll have to build into the job any mechanism to pick up where you left off, something not all that trivial. It involves keeping track of your last commit point and then leveraging that after a restart to skip all of the committed rows, typically by a transformer constraint that only lets rows past after the last commit point has been reached. And doing this in a multi-node PX job complicates things.

As for the commit level, what is your target database? What write method are you using? Can you switch to the ODBC Connector stage? It supports the concept of Transaction sizes, from what I recall.

[kannan_ayyanar]

DB2 database and write mode is insert..

i can switch to ODBC connector stage... that is not a problem....main concern is the record level commit... is there any option in ODBC connector stage...?

[chulett]

It should, yes, hence the recommendation. Transaction Size = Commit Level.