Data Types having Dimensions

In the LN Studio, repeatable attributes are modeled by setting the ‘dimensions’ for a data type. Repeatable attributes are attributes that can occur multiple times within their parent. This concept is comparable to array columns as used in the LN data model.

In a business interface implementation, the following can be modeled for LN:

• One-to-one mapping:

  Use an attribute having n dimensions, use an array column having the same number dimensions, use an attribute implementation having the same number of dimensions. In theory an attribute (and attribute implementation) can have more dimensions if it is input-only. And an attribute (and attribute implementation) can have fewer dimensions if it is output-only. But these situations are unsupported in a one-to-one mapping. In such cases a calculation (see below) must be used instead.

• Element mapping:

  Use an array column (data type of table column has multiple dimensions) and map one or more elements to attribute implementations. This is done through the ‘Additional Instructions’ property. For example, a table has a column prcs (Prices). Three attributes are defined, being costPrice, localPrice and cataloguePrice. These attributes must be mapped to the first three array elements of the prcs column. In that case, map costPrice to column prcs and in the Additional Instruction for the attribute implementation, fill in “1”. For localPrice fill in “2” and for cataloguePrice use “3”. The attributes and attribute implementations use a data type with no dimensions. Only the table column will use dimensions.

• Using a calculation:
  • On get hook, calculated attribute implementation has dimensions, and used attributes have no dimensions.
  • On get hook, calculated attribute implementation has dimensions, and used attributes have one or more dimensions.
  • On get hook, calculated attribute implementation has no dimensions, and one or more used attributes have multiple dimensions
  • On set hook, attribute implementation has multiple dimensions, and used attributes have no dimensions.
  • On get hook, attribute implementation has multiple dimensions, and used attributes have one or more dimensions.
  • On set hook, attribute implementation has no dimensions, and one or more used attributes have multiple dimensions

It is not possible to map a repeatable attribute directly to separate rows of a table for the component. For example, component Order maps to the Order table. Attribute Order.note having multiple dimensions maps to the rows of an OrderNotes table. In such case, the note attribute implementation must be implemented using an on get and on set hook, which reads and updates the table (through its DAL).

The same domain can be used in LN for table fields with or without dimensions. In that case multiple datatypes are needed, one for every combination of domain (native data type) and dimensions value. Each data type must have a unique name, but the native data type will then be the same. For example:

When using repeatable attributes, the constraint are applicable for LN:

• In case of a ‘one-to-one mapping’, the value of the dimensions property must match.
• In case of an ‘element mapping’, the specified element must exist. In other words, the element must be greater than 0 and less than or equal to the number of elements for the array column.
• If filled, the dimension must always contain a number. Other values (including for example multi-dimensional arrays such as “10, 20”) are not allowed.
• A dimension can only be used for a leaf attribute. So it cannot be used for a ‘complex’ data type or for an attribute group.
• A qualifier cannot be repeatable.

A repeatable attribute can be defined as dynamic by setting the dimension to 0. In that case the number of repeatable elements does not need to be known at beforehand. This is useful for processing incoming BODs.

if i.optionalNote.iSet then
		o.note(1,1) = i.optionalNote
	else
		io.note.isSet(1) = false
	endif
	o.note(1,2) = i.extraNotes(1,1)
	o.note(1,3) = i.extraNotes(1,2)
	o.note(1,4) = str$(i.numbers(1))
	o.note(1,5) = str$(i.numbers(2))
	return(0) | OK

if not i.code.isSet(1) or not i.code.isSet(2) then
		io.cancel = true
	endif
	return(0) | OK

	#pragma used dll oppmmmdll0001
	return( ppmmmdll0001.calculate.quantites.and.add.notes(i.item, i.item.isSet, 
									i.description(1,1), i.description.isSet(1),
									i.description(1,2), i.description.isSet(2),
									io.quantity(1), io.quantity.isSet(1),
									io.quantity(2), io.quantity.isSet(2),
									o.note(1,1), io.note.isSet(1),
									o.note(1,2), io.note.isSet(2)) )