commit transaction

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

karthi_gana
Premium Member
Premium Member
Posts: 729
Joined: Tue Apr 28, 2009 10:49 pm

commit transaction

Post by karthi_gana »

All,

when we use "Server job", i know there is an option to mention "Rows Per Transaction" to commit the transaction.

But i don't see that option in "Parallel job". is it because parallel job use the BULK COPY method?
Karthik
manoj_23sakthi
Participant
Posts: 47
Joined: Tue Feb 23, 2010 12:16 am
Location: CHENNAI

Re: commit transaction

Post by manoj_23sakthi »

could u please give the correct information about the stage?
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

I doubt that U could give the correct information, because U works in Singapore for a different company than does karthi_gana.

The second person personal pronoun in English is spelled "you".
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

Well, someone let us know what stage we're talking about here. :wink:
-craig

"You can never have too many knives" -- Logan Nine Fingers
karthi_gana
Premium Member
Premium Member
Posts: 729
Joined: Tue Apr 28, 2009 10:49 pm

Post by karthi_gana »

ODBC Stage - in Parallel
Karthik
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

Bulk loading is non-transactional, so that COMMIT is not relevant.

But the ODBC Enterprise stage also allows you to use Upsert method. Then transactions are relevant, and the number of rows per transaction can be controlled either by an environment variable or by a stage property that becomes enabled when Upsert method is selected.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
qt_ky
Premium Member
Premium Member
Posts: 2895
Joined: Wed Aug 03, 2011 6:16 am
Location: USA

Post by qt_ky »

Which ODBC stage type?
Choose a job you love, and you will never have to work a day in your life. - Confucius
karthi_gana
Premium Member
Premium Member
Posts: 729
Joined: Tue Apr 28, 2009 10:49 pm

Post by karthi_gana »

can you tell me which stage property control this?
Karthik
qt_ky
Premium Member
Premium Member
Posts: 2895
Joined: Wed Aug 03, 2011 6:16 am
Location: USA

Post by qt_ky »

karthi_gana wrote:can you tell me which stage property control this?
Yes. You go first...
qt_ky wrote:Which ODBC stage type?
Choose a job you love, and you will never have to work a day in your life. - Confucius
karthi_gana
Premium Member
Premium Member
Posts: 729
Joined: Tue Apr 28, 2009 10:49 pm

Post by karthi_gana »

which ODBC type?

ODBC Enterprise ! ?

or

Parallel Job - ODBC Enterprise to connect SQL Server
Karthik
qt_ky
Premium Member
Premium Member
Posts: 2895
Joined: Wed Aug 03, 2011 6:16 am
Location: USA

Post by qt_ky »

Bulk load is not a default setting for ODBC stages.

Note: ODBC Connector is the preferred stage type for ODBC.

For the ODBC Enterprise stage, the default Write Method = Write and default Write Mode = Append. To override the default commit interval:

Go to Options, select Insert Array Size. The default is 2000. Set it to 1.

Go to Options again, select Row Commit Interval. Set it to some number.

Initially you won't find the Row Commit Interval setting. You'll only find it after you set Insert Array Size = 1.

The default Row Commit Interval is equal to the Array Size, whatever you have set it to.
Choose a job you love, and you will never have to work a day in your life. - Confucius
karthi_gana
Premium Member
Premium Member
Posts: 729
Joined: Tue Apr 28, 2009 10:49 pm

Post by karthi_gana »

Note: ODBC Connector is the preferred stage type for ODBC.
can you tell me why?
Karthik
karthi_gana
Premium Member
Premium Member
Posts: 729
Joined: Tue Apr 28, 2009 10:49 pm

Post by karthi_gana »

I have also seen APT_RDBMS_COMMIT_ROWS environment variable in the below link.

http://etl-tools.info/infosphere-datast ... iables.htm

I don't see this variable in my project. !!! ???

I can see $APT_ORAUPSERT_COMMIT_ROW_INTERVAL in my project. The dafault value is 5000.

We are using SQL 2008. So i am sure i can't use this variable.
Karthik
karthi_gana
Premium Member
Premium Member
Posts: 729
Joined: Tue Apr 28, 2009 10:49 pm

Post by karthi_gana »

I have tried to learn about ODBC stage and its option (like Insert Array Size, Fetch Array Size, Row Commit Interval etc.,). But i don't see this stage in any manual. If anybody knows where i can found more info on this, please let me know.

Meanwhile can you explain what is the difference between 'Insert Array Size' and 'Row Commit Interval' ?
Karthik
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

karthi_gana wrote:
Note: ODBC Connector is the preferred stage type for ODBC.
can you tell me why?
Connectors are the way of the future and are thus preferred across the board. All other stages are deprecated, which the documentation clearly states.
-craig

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