Talking about partitioning
When you need to load massive amounts of data to your database, partitioning might be another optimization option. But you really should be confronted with massive amounts of data when you start considering partitioning.
Do you remember the math of the CCI and why it will only perform when you load around 63 to 100 million rows to your database (see the preceding section, Understanding CCI)? Now, you need to add another factor to this equation: the number of partitions that you are planning for your database.
Let's assume that you want to have one partition for every month (the most typical usage of partitions) in your table, and you plan to load data for 5 years to your database. This will add another 60 as a factor to your preceding term: 60 distributions x 1,048,578 rows per distribution x 60 months in the database. This results in 3,774,880,800 rows that your table needs to hold as a minimum in order for the CCI to be built over all the partitions...