In all of the Moving Function calculation examples, it is assumed that the Moving Function calculation proceeds in a downward direction for each Computed Item column. That is to say, the Moving Function calculation window progresses downward, one row at a time, for each subsequent row of displayed fact data. This downward movement is evident in the Table section data (which has been used in the examples shown above) since, in a Table section, each new instance of fact data can only be represented as a new row.
Other sections, however, such as Pivot and Chart allow the possibility for fact data to be represented in more than one directional axis. In the case of Pivot and Chart sections, therefore, you can specify the direction of Moving Function calculation. To illustrate the effect of changing the direction of Moving Function calculation, consider the previously shown Table data:
Adding a new Computed Item column to display a 3-Month Moving Difference would yield:
Consider, however, if the "Month" and "Sales" data were oriented as follows (as could be done in a Pivot section):
If a new Computed Item were added, to represent a 3-Month Moving Difference, and the Moving Function calculation direction was still down each column, this would result in null values being represented for the Computed Item (as below) since there would only be one fact value per column (this would not be enough fact occurrences to satisfy a 3-term Moving Difference).
If, instead, the Moving Function calculation direction were specified to be along each row (i.e. horizontally from right to left) the Moving Difference Computed Item would yield the "expected" results since the source "Sales" Fact data cells would be examined in their "correct" left-to-right sequence. That is, the following display would result:
| Month | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|