See this article in Bi4 NewGen
Text and Numeric Filters
Bi4Cloud provides filters to allow you to segment your data. These filters can operate text values in the data and on the numeric values of metrics.
The filter panel displays when you press the filter icon and the panel below is displayed. The top section are the text filters and at the bottom are the numeric filters.
You create a Numeric filter by clicking Add a Numeric Filter and choosing the type of metric to filter.
Here is a sample of the Numeric filters from the Sales Vs All Inventory mapping.
Note that there are two styles of metric filter. Ones based upon filtering at the Source of the data and ones that filters based upon the Results query.
The Source data is the lowest level tat a report goes to and in Sales Vs All Inventory this level is the Item and it's Location which is shown below. This example is filter for 4 Items.
No Numeric Filters Applied
The above shows a column for Last Sold Wks and On Hand Stock for each location, grouped by Item. Below we Analyse by is Item and we see the On Hand as the sum of the Item locations but notice the Last Sold Wks in the minimum of the Item Locations Last Sold Wks. It's not the sum of the Last Sold Wks because it needs to be the minimum of the last weeks when it was sold. That is what the arrows highlight in the above shot.
Using Numeric Filters
Let assume you want to retire ( set InActive ) all stock that you have not sold for the last 6 months (27 weeks) and that has 0 stock on hand.
Let's choose the qty_onhand (source) = 0 and apply this in Analyse by InvoiceStockLocation, then only locations with zero stock are displayed.
All Locations with zero stock are displayed - notice the on hand sums for the item sub-total are zero because the locations with non-zero stock have been filtered out.
Now add in weeks_since_last_sold (source) > 27
and locations that have not sold for 27 weeks are filtered for.
Now switch to Analyse by Item
Great, but this, however, is the wrong result if we intend to use this to retire stock because with reference to the original screenshot No Numeric Filters Applied we know there are locations with stock that have sales < 27 weeks ago but these are excluded because the locations have non-zero stock.
This is where the Result filters become vital. These filter on the results of the Analyse by dimension. Lets change the filters to the equivalent Result filters.
If we applied this to the Analyse by InvoiceStockLocation dimension then the result would be the same as Source filters because that is the lowest or source dimension. However when these filters are applied to the Analyse by Item dimension the aggregate values for the Items are used for the filters. This shows there are no Items that have On Hand = 0 and Last Sold Wks > 27 which is the corrrect result.
Generally it is best to use the Result Numeric filters because these filter on the dimension being displayed.
Source filters do have their uses however. An example would be to eliminate give-away or small value lines Sales Analysis. The Analyse Sales Mapping has the lowest transaction level or source level of the Invoice Line Detail. So to eliminate the small value lines you can use a source numeric filter on the invoice line amount < $1