Release 22nd January 2025
Change Control note --version 2.1.0
Overview
This release introduced introduced two changes, to enhance error reporting when binding an element that fails a horizon filter, the second add functionality to simplify schema evolution. In addition to changes to Hiperspace, packages references have been upgraded to the latest versions.
Horizon Filters
Horizon filter provide a conditional view that divided a subspace with elements that are visible in the hiperspace, and those that are not (over the horizon). This is great for reading from hiperspace because it allows elements that are not valid in the current context (e.g. changes that have not been approved) or the user does not have access. The Horizon filters are also applied when binding an element to hiperspace, since you should not be able to add anthing that you could not see.
Hiperspace horizon is contextual, to allow the same code to be used for initial seeding of data and logging errors for later analysis. This change assists with validation by adding the option of an error message when the horizon is used for data validation.
new Horizon<Element_Type> (element => element.Valid == true, "element only")' a horizon predicate with the element only new Horizon<Element_Type> ((element, read) => element.Valid == true || read, "element only")' a horizon predicate with a flag to indicate a read operation
new Horizon<Element_Type> ((element, context, read) => element.Valid == true || read, "element only")' a horizon predicate with a context parameter and a flag to indicate a read operation new Horizon<Element_Type> ((element, context, user, read) => element.Valid == true || user?.IsInRole("BULK"), "element only")' a horizon predicate with a context parameter, user IPrinciple and a flag to indicate a read operation
Auto-edit of #id alias in HiLang schema
If a #id is not added to the definition of an {element, key, value, extent, index} an id is auto allocated during the compilation of a .hilang file. It is best practice to add #id field to simplify the evolution of a schema since the internal #id can never be changed (to a different data type) or reused for a different purpose since it wil invalidate historic views of data. This can be a problem when {element, key, value, extent, index} are dropped and the #id is then used for a later addition. To simplify maintenance the directive %ids; has been added to optionally add the generated #id to the source file
the source entity Customer (Id : Int32) {Name :String} [SameName = Customer(Name = Name]; defines an entity Customer with a key Id and value Name and an index SameName that can be used to find Customers or to apply horizon validation 'Horizon<Customer>("Duplicate Name", (customer, read) => customer.SameName == null || read).
When the %ids; directive is added, the the source is auto edited to entity Customer #1 (Id : Int32 #1) {Name : String #2} [SameName = Customer(Name = Name) #2]; to indicate that Customer has #1 id and CustomerSameNameIndex has #2 (stored items have unique #id, but SameName is not stored with customer, so does not need to be unique within the element)
If we decide to break Name into First_Name and Family_Name we'd change the entity to
entity Customer #1 (Id : Int32 #1)
{First_Name : String #3, Family_Name : String #4}
[Name = First_Name + " " + Family_Name, SameName #3 = Customer( Family_Name = Family_Name, First_Name = First_Name];
Adding %ids; directive simplifies the process of avoiding clashes that will prevent the hiperspace being opened with updated code.
DeltaIndex
The @DeltaIndex property of an entity provides indexes access to elements that have changed since a date-time stamp. The property has been updated to include the option to define the #id for the index by adding a parameter @DeltaIndex(4)
CubeFact
The @CubeFact property of an element creates a fact element (the code @CubeFact entity Account adds an additional Account_Fact table to the hiperspace for stored historic aggregates of @CubeMeasure values).
The property has been updated to include the #id of the fact table @CubeFact(5) indicates that the fact table has #id of 5.
%ids; is optional since #id's will be allocated (during compilation) for intermediate elements that will be pruned before generation resulting in gaps in the generated #id numbers