OLTP or OLAP
OLTP or OLAP
But i am at a deciding point where i have to know: is olap actually the right system to use here?
There is the following situation:
I have a table (~400.000 rows) (no data mart or data warehouse) with the following information:
ID HOMEPAGE DATE_ARRIVAL TIME_ARRIVAL DATE_LEAVING TIME_LEAVING
-> This table shows the user-visits on a specific homepage (when they arrive and leave) and is not changeable.
So the mission is to analyse this. For example by drawing a diagram at the front-end that shows the visits during the day xy on homepage z: Then there is an x-axis (Time) and an y-axis (Visitors). As there are many visitors on one day i have to cluster them into groups so that i can draw a nice graph. Let's say i cluster them every half hour. So i sum up all visitors that are on the page from 15:00 to 15:30.
Now i can do many things with this information for example I can draw the same diagram also with the informations of a week/month/year where i have to use an average value for the clusters (sum(visits)/days) . Then there is also the possibility to distinguish the graph between months/years, homepages and so on.
As there are just monthly updates on the original table and i need a good response rate (< 5 sec) i thought OLAP will be the way to go. But the dimensions Homepage, Time, Date lead me to a fact table with 30 millions rows.
Do you agree or disagree on using OLAP here?