Uniqueness Verification in Transformer

[sambias]

I have some questions about Uniqueness Verification flag in Data Source properties in Transformer:


If I have a data source, .csv file for example, used only for fact table, with no duplicate records, I could set the uniqueness flag at “Maximize data access speed” without incurring in wrong summarization? I use other .csv file for dimensions.


I have done some test and I have noticed that “Maximize data access speed” decrease the fact table time reading phase about 30%.

[Arsenal]

It really depends upon the source data. It's not just about duplicate data, it's also about what sort of data you're using to report on. For example, a city name might occur in 2 different states in a country, so you need to let Transformer know that by setting uniqueness at the correct level. If in this case, you set the City to be unique but still choose to go with Maximize Data speed, Transformer will not inform you of the uniqueness violation and you will have incorrect data to report on.

If you're very, very sure that there won't be uniqueness violation, then you can use Maximize Data Speed.

[jive]

Hi,
Did the maximized access.... his not about how the cube will be build. If I remember correctly higher the access speed is set in transformer affect directly the time of build and the sized of the cube.

I'm right or completely out of context here