Too much of inner joins-ETL Strategy to load into Dimension

Hi,
We have situation where the tables are not directly linked with each other. To select one value many tables has to be linked. For eg. to select column c1 from table T1, we have to put inner join between t2,t3 and t4 with T!. Same way, if we need to select c2 from T1, we have inner join with T2,T5 and outer-join with T6. Which means that, there are no direct relationships between the data and we need to take the T2,T3,T4,T5 & T6. Morever, possibility of missing data possible due to outer joins. Will this not result into Snow-flaking in the star schema. In this kind of scenario, how to create dimension and fact without snow-flaking?

-RBS

Let me see if I understand the question...

1. You have a source where you need to join a bunch of tables together to get the data you need.

2. Sometimes there isn't data in some of the tables.

3. You need to load it into a dimension.

So, what is the problem? Why do you feel there is a need to snowflake?

ngalemmo wrote:Let me see if I understand the question...

1. You have a source where you need to join a bunch of tables together to get the data you need.

2. Sometimes there isn't data in some of the tables.

3. You need to load it into a dimension.

So, what is the problem? Why do you feel there is a need to snowflake?

The problem here is the data in the joining tables are candidates for dimension.
So if the data has dependency on the other dimension, will it not create snow flake?

No. Relationships between dimensions are based on the fact table. Dimensions do not relate to each other.

In other words, if a customer belongs to a sales region, and both customer and sales region are their own dimensions, the fact should have FKs to both customer and sales region if it makes sense for the facts.