Informatica Connector. Why not read from database table?

[#1]

Hi!

I'm searching for arguments for implementing the Qlikview Informatica Connector. What are the advantages of creating Qlikview targets in Informatica versus making Qlikview reading from database table with an ODBC connection.

Anyone have practical experience?

Anyone have a usercase on the Informatica Connector?

Regards

Mathilde Andersen

[rbecher]

Hi Mathilde,

a QVX file can be loaded much faster into QlikView than records from a database table. Mostly databases are on a separate box so you're facing a network bottle-neck.

So, if you have to load a lot of records, lets say more than a million, QVX is an option.

We will probably use the Informatica connector also for the reason that the target (QVX file) could be provided earlier for the QlikView load because database write performance is also a bottle-neck in the whole ETL-process.

- Ralf

[#3]

Hi Ralf.

Thanks for the reply.

What you are saying is that Informatica will spend less time writing a QVX than Qlikview is able to read a database table?

Would you prefer to have database, Informatica and QlikView on the same box?

- Mathilde

[rbecher]

Well, it depends. In my experience it could be that creating a QVX file with Informatica and loading this into QlikView is much faster than writing all records into a table (where you probably have data types, constraints like not null and PK etc.) and loading from a table via Select. But, it's depending from so many parameters. The suggestion is to measure the two different approaches.

In order to prevent a network bottle-neck we suggest to run PowerCenter on the database instance. But for memory and performance reason, the QlikView server should always run on a separate machine!

- Ralf