I have a question about the lookup files that get exported with the clickstream data. We export data every hour, and in each export I get a set of lookup files.
The lookup files contains two fields - an id and a value. The lookups are essentially dimension tables, the main clickstream data set is the fact table.
As I make each export available in our data warehouse I have a choice - depending on what assumptions I want to make about the lookup tables:
There is nothing that I can see in the documentation. This biases me towards the "Denormalize and store' strategy. However this is much more complex process to build and requires considerably more storage. My initial checks seem to confirm that I will be OK to store as-is and assume that the lookup tables will only ever get new ID's added and old ID's will not be recycled or removed.
I would love confirmation - what is the Adobe Clickstream Data Model - so that I can make the best choice?
Thanks,
Matt
Solved! Go to Solution.
I think one of the reasons why they're separated like this is so you have that choice. I will say that I've only seen lookup table data added to, and never removed. The removal of lookup data would have an adverse effect on historical data collected.
If denormalizing and storing is the better option for you, then I don't anticipate any issues with the clickstream lookup tables interfering.
Views
Replies
Total Likes
I think one of the reasons why they're separated like this is so you have that choice. I will say that I've only seen lookup table data added to, and never removed. The removal of lookup data would have an adverse effect on historical data collected.
If denormalizing and storing is the better option for you, then I don't anticipate any issues with the clickstream lookup tables interfering.
Views
Replies
Total Likes