Contents x
      Profile - Dimension
      • 2 Minutes to read
      • Print
      • Dark
        Light
      • PDF
      Contents

      Profile - Dimension

      • 2 Minutes to read
      • Print
      • Dark
        Light
      • PDF
      Article summary

      Dimension

      The dimension determines the category being used to group data. For example, Months, Sales regions, Cities, Products, Ages, Sizes, Colors…


      A dimension is usually a database field – it can be thought of as a column of data in a spreadsheet.

      Any discrete field can be added to a profile as a dimension.    

      A dimension can be thought of as a rows field in an excel pivot table. 

      Dropping a field onto the Dimension Box will automatically calculate and display a profile:


      Right-clicking a dimension will display the Dimension Menu:

      Remove...Removes the currently selected dimension from the calculation
      PivotApplies the selected dimension as a columns dimension and creates a crosstab.  If a pivot has been applied, the dimension will be displayed in a different color in the dimension box:
      Note:
      Only 1 dimension can be specified as a column dimension at a time
      Clear PivotReturns the selected column dimension to a row dimension
      Navigate ToHighlights the selected dimension in the Database Tree


      Discrete or Continuous

      A discrete field is a field that has a limited number of discrete values – up to a maximum of TODO - as opposed to continuous.   DateTime fields are always continuous fields, and key fields are often continuous fields. 

      Adding a continuous field as a dimension to a profile will generate a warning message saying that it cannot be calculated. 

      For example, the field Months is a discrete field with 12 discrete values.

      The field Countries is a discrete field with up to 195 discrete values. 

      The field Session_Start_Time is a continuous DateTime field with an unspecified number of values.


      Profiling DateTime fields
      To use DateTime fields in a profile report, first categorise them into a discrete grouping using one of the DateTime functions - for example, YearMonth, WeekOf...etc   See Engineering | Functions | DateTime Functions for more details 


      Single and Multi Dimensions

      Profiles can be single-dimensioned, or multi-dimensioned.   

      A single-dimension profile has 1 dimension and simply counts the number of records that have a particular value:

      A multi-dimensioned profile has more than 1 dimension and takes the counts for a single dimension and splits them out by another dimension.

      A multi-dimensioned profile that has a pivot applied is called a Crosstab.  To add a pivot to a profile, Right-Click the dimension and choose Pivot:

      Dimensions are analogous to the Rows Fields and Columns Fields of an excel pivot table.    

      Dimensions can be raw or pivoted.      If raw, a dimension acts like a rows field.   If pivoted, it acts like a columns field

      Order and Sorting of Dimensions

      The order in which the dimensions are calculated will affect both display and results.    Dimension order can be changed by moving dimensions left and right.   Sort order will also affect the display - this can be controlled by clicking on the result grid, or by using the sort options in the micro-toolbar.


      Dimensions and Tables

      The dimensions in a profile must all come from the same table or from tables which are linked.    The results returned by the profile will always refer back to a single table.   

      The table of the first dimension is the default resolution table – i.e., determines which records are being counted.     If the first dimension is [Product.Product Group], results will refer to product records.   If the first dimension is [Customer.Age] the results will refer to Customer Records. 

      Note
      It can be difficult to accurately interpret results that are returned from profiles that are using dimensions from different tables

       Further control over what is counted is achieved by using the Resolve At function. 


      Was this article helpful?
      What's Next