Introduction to roll up columns
Rollup columns perform aggregate calculations on values that are stored in a column in a related table across a one-to-many relationship. Rollup columns can calculate sum, count, minimum, maximum, and average values in a fashion that is similar to how these same functions are used within Microsoft Excel. Rollup columns are read-only and can calculate values based on a system account with full access to all rows in a related table, so they might calculate values based on rows that a current user can't access due to security settings that prevent the user from access.
Rollup columns are calculated by Microsoft Dataverse and the value of the rollup calculation is stored within the column in the table, meaning that rollup columns can be used for sorting, searching, and filtering like any other column.
When you configure a rollup column, Dataverse starts two recurring jobs. These two jobs are a Mass calculate rollup and a Calculate rollup.
Calculate rollup
Rollup calculations are calculated asynchronously (meaning not in real time) by using a scheduled system job, Calculate rollup, in Dataverse. Rollup calculations run once an hour by default, but an administrator can schedule them to run as often as needed by accessing System Jobs from the Power Platform admin center.
The Calculate rollup job is an incremental calculation of the rollup column and only update the column to include created, updated or deleted rows.
To change the scheduling of a rollup system job, select the environment in the Power Platform admin center and then go to Settings > Audit and logs > System jobs. In the System Jobs window, select the All Rollup Field Calculation Jobs view and select the rollup job for the column you want to change, select Modify Recurrence under Actions and update the schedule.
Mass calculate rollup
The Mass Calculate Rollup Column occurs, by default, 12 hours after you configure the rollup column (to coincide with nonoperational hours of the environment). This job calculates the rollup column for all existing rows that contain the column.
For more information on these rollup jobs, see Define rollup columns that aggregate value.