Page 1 of 1
better option for informix database
Posted: Sun Feb 06, 2005 8:38 pm
by rajiivnb
Hi,
My source is a informix database .Which stage could be easy and better for extracting the source if ODBC or informix Stage.
Regards,
Raji
Posted: Sun Feb 06, 2005 8:58 pm
by chulett
I don't have any hands-on time with the Informix stage, but in my experience it's generally better to use a 'native' stage over something generic.
On the other hand, an ODBC wire driver can be used when you don't have the client software installed on the DataStage server.
Posted: Sun Feb 06, 2005 9:03 pm
by rajiivnb
Hi Craig,
Is performance increased while using a native stage .
Rajii
Posted: Sun Feb 06, 2005 9:14 pm
by chulett
One would certainly hope so.
Plus you should have more control over the various options available when working with your database of choice.
Easy enough to test. Build two jobs, one with an ODBC stage and one with an Informix stage as the source. Have them select everything from the same large table. Write the output to a sequential file but put a constraint of
@FALSE in the output link. This will cause nothing to actually be written to the file and will show you the fastest speed at which you can pull data from the database on your system. Compare and contrast.
Posted: Mon Feb 07, 2005 3:38 am
by Sainath.Srinivasan
You can almost always expect a performance improvement when using a native stage. This is because the native stage knows more about the db and its tuning than the ODBC mechanism.
Also any new version of db will be accompanied with the corresponding native drivers and libraries to include new feature's benefits, improvements and bug-fixes.
The main advantage of using an ODBC stage is that you can change it to point to a new rdbms quickly as most configurations are outside DS developers view.
Posted: Mon Feb 07, 2005 7:49 am
by chulett
If you get into a situation where you feel you may need to change the underlying RDBMS, there always the
DRS or Dynamic Relational Stage. It was built for this very reason and allows you to shift from various native drivers to ODBC by setting a parameter. Not sure if it supports Informix natively, however.
More of an FYI to the original poster than anything...
Posted: Mon Feb 07, 2005 8:25 am
by kduke
Informix has an unload option which will create a sequencial file fast. Their ODBC drivers are slow.
I do not remember the syntax for unloading a table.
Posted: Mon Feb 07, 2005 8:30 am
by Sainath.Srinivasan