Data for the current selection is calculated only when it is requested by the user. The selections are used to filter the underlying dataset before the metric values are aggregated at the level of the Grid/Graph that is displayed in the document. If the source attribute is not included in the Grid/Graph, the metric values from all the selected elements are aggregated and shown at the level specified in the Grid/Graph.
Data for each available item in the selector is calculated in advance when the document is first displayed. Selections made while viewing the document are used to determine which slices of data are combined and shown in the Grid/Graph. Even if the source attribute is not included in the Grid/Graph, the data is still sliced at the level of the source attribute, and therefore the metric values from multiple selected items are not added together. Instead, the data for each selected element is shown separately in the Grid/Graph, the same as if the source attribute had been included in the Grid/Graph.
The dataset report of a document contains Region, Year, and the Revenue metric. A Grid/Graph displays Year and Revenue only, and is targeted by a selector with Region as its source. The selector is defined to slice the data. When Central is selected, three rows, one for each year, are displayed, as shown below:
If you select Mid-Atlantic as well as Central, six rows are displayed, two for each year, as shown below
This occurs because the selector slices the data by region before the user selections are made, and cannot aggregate the slices for multiple regions.
If you change the selector to filter rather than slice the data, the yearly revenue is aggregated across the selected regions. The yearly revenue is calculated by adding the Central and Mid-Atlantic values for each year, and only one row for each year is displayed in the Grid/Graph, as shown below:
Other important differences between filtering and slicing selectors are described below:
1) Slicing allows the total to be displayed as an item in the selector. A filtering selector does not display the total as a selector item.
2) Slicing allows you to specify that the selector automatically uses a default selection when other changes in the document cause the selection made by the user to return no data.
3) If a selector is sliced, you can define the current state, which determines how the target is displayed when the document is executed. The target can display all the selector items, a specific number of the first items in the selector, or a specific number of the last items in the selector. If a selector is filtered, you can define the current state as unset only, which displays all the selector items.
When a document is viewed off-line (exported to Flash, in a subscription, or in MicroStrategy Office):
1) If a selector is applied using filtering, only the data for the current selections are included in the document. An off-line user cannot change the selector and update the target.
2) If a selector is applied using slicing, all the slices, and therefore all the data, are included in the document. An off-line user can change the selector and update the target.
3) To continue with the yearly regional revenue document described above, the selector is applied as a filter. Only Central is selected, and the document is exported to an MHT file to be used off-line, without using MicroStrategy. The MHT file contains only the data for Central, and no other selections can be made.
4) If the selector is applied as a slice instead, all the data is sliced and included in the MHT file. Even if only the Central region is selected when the document is exported, you can use the selector in the MHT file and display other regions.