Group-by should support grouping rows' logic based on predicates

[Seddryck]

The current implementation of group-by supports grouping by distinct values and equality (Two rows belong to the same group if they share the same value for the columns part of the group-by statement).

It should be possible to also group rows based on one or a combination of predicates defining one or more groups.

<group-by>
  <case>
     <predicate operand="colValue">
         <within-range>[0;1000[</within-range>
     </predicate>
  </case>
  <case>
     <predicate operand="colValue">
         <more-than>1000</more-than>
     </predicate>
  </case>
</group-by>

at no moment it should be mandatory that the predicates operate on a unique operand. This is valid …

<group-by>
  <case>
     <predicate operand="colValue">
         <within-range>[0;1000[</within-range>
     </predicate>
  </case>
  <case>
     <combination operator="and">
       <predicate operand="otherValue">
           <lower-case/>
       </predicate>
       <predicate operand="myValue">
           <any-of>
              <item>Saturday</item>
              <item>Sunday</item>
            </any-of>
       </predicate>
     <combination>
  </case>
</group-by>

All rows not validating any predicate will be part of an additional group automatically created.

[Seddryck]

Beta available at NBi 1.22.0-beta.69