• Creating Named and Dynamic Collections
• Decision Tables and Collections
• Processing Collections With Decision Tables
• Sep 2015 Collections of Cars
• Tutorial Testing with a collection of inputs

There are three basic collections.

1. The basic collection is often called a bag. It stores objects with no ordering of the objects and no restrictions on them.
2. Another unstructured collection is a set where repeated objects are not permitted: it holds at most one copy of each item. A set is often from a predefined universe.
3. A collection where there is an ordering is often called a list. Specific examples include an array, a vector and a sequence.

• Collections are a bunch of objects that can be named and referred to.
• Different types of object can be grouped into a single collection.
• Objects in a collection can be basic data types, or structured data types or even collections.
• Special operations are available for processing collections
• The members of a collection do not even have to be of the same type. example ["red","red",3.24,"2022/12/31",true,[1,2,["a","b",5]],"red"] is a collection

• Sets are collections of objects that contain no duplicates
• The objects in a set are called its elements or members.
• The elements in a set can be any types of objects, including sets!
• The members of a set do not even have to be of the same type. Example ["red",3.24,"2022/12/31",true,[1,2,["a","b",5]]]is a set

• Lists are collections of objects that have been arranged into a specific sequence

• A collection can be sequenced on one or more attributes using the sort function
• An operation can return any data type including a collection.

Multiple collections can be joined and filtered.

Example: Generating cross products Suppose you have a collection of persons such as these:

{
  "persons": [
  {"name":"Tom","age":57,"salary":100000,"gender":"male"},
  {"name":"Dick","age":26,"salary":80000,"gender":"male"},
  {"name":"Harry","age":37,"salary":160000,"gender":"male"},
  {"name":"Jane","age":42,"salary":120000,"gender":"female"},
  {"name":"Mary","age":26,"salary":180000,"gender":"female"},
  {"name":"Ruth","age":76,"salary":180000,"gender":"female"}
  ]
}

NOTE
avgAge={p in Person where age in [20..65],gender='male'}.sortedBy(age).first(3).age.avg can be expressed as a series of statements:

1. adultMales={p in Person where age in [20..65],gender='male'}
2. adultMalesAscendingAge=adultMales.sortedBy(age)
3. youngestThreeAdultMales=adultMalesAscendingAge.first(3)
4. avgAge=youngestThreeAdultMales.age.avg

Person.iterate() Eg {p in Person}.iterate(p.net=p.gross-p.taxes) updates p.net for all persons

size, iterator(), add() or +=, remove() or -=, clear

Collection must be sorted eg sortedCollection=collection.sortedBy(age)

Assume A={red,green,blue} and B={red,yellow,blue,purple} where A and D are sets of instances of the same type

Each should implement a common interface Common attributes (if any) must be of the same type. Common methods (if any) may have different implementations but must return the same data type

[list of any valid expression involving attributes of the collection that evaluates to boolean]

{p in Person where age>65} just those instances where age >65.

Person is equivalent to Person[] and Person[*] meaning every instance of the class with no filtering. Person[n] any n arbitrary instances (can't think why we'd need this yet)

{p in Person where age>65,gender='male'} allows the filters to be applied in any order (perhaps even in parallel with the final result being the intersection of the two collections. This is equivalent to {{p in Person where age>65} where gender='male'} which requires the age filter to be applied first and then the gender filter is applied to the results. Not sure if it makes a huge difference