The interval of your data can significantly impact the outcome of your calculation for specific rollup intervals. This is outlined in the example shown below.
The example shows the same data (electricity & production) for the same date range (Today) but at different intervals of time. The table on the left is shown in 5-minute intervals and the table on the right is shown at 10-minute intervals.
The calculations for electricity/production and electricity*Production are then calculated for each row as shown.
The table below shows the statistics for each of the grids above.
A simple summation of the datasource columns (electricity & production) at both intervals (5 min and 10 min) give the same results. This is the expected results and tells us that the same amount of production and the same amount of electricity was consumed for the date range.
However, we notice that there is a discrepancy between the results of other statistics such as average, min, max, etc. These statistics are single value outputs of the data shown in the data grid. This is due to the different time intervals and the inputs into each of the functions.
I often might want to simply know what electricity/production or electricity*Production is for the entire date range. So today the total electricity consumed is divided by the total production. The output of these calculations is shown above at 2.18 and 264.00 respectively for each of the time intervals.
To show this a user must create a Single Value in their workbook.
<How to Create a Single Value>
This calculation looks like the following:
SUM(electricity for date range) = 24
SUM (Production for date range) = 11
24/11 = 2.18
As opposed to the following, which applies the calculation at each row before returning the calculation.
electricity/production for time interval 1 = 1.50
electricity/production for time interval 2 = 6.00
electricity/production for time interval 3 = 2.33
electricity/production for time interval 4 = 2.00
SUM (1.50, 6.00, 2.33, 2.00) = 11.83
These calculations are both correct but depend on what sort of output is required.
When I require a single value the summation of the input columns is carried out first.