I've been doing some testing, and it appears you do not need a record with a root hierarchyid.
For example, normally you would one root node (level 1) and multiple childen, but you can skip the root node, having no root records, just records that start at level 2:
//table schema
CREATE TABLE [Entity](
[ID] [int] IDENTITY(1,1) NOT NULL,
[Name] [varchar](50) NOT NULL
[Hierarchy] [hierarchyid] NOT NULL,
CONSTRAINT [PK_Entity] PRIMARY KEY CLUSTERED
(
[ID] ASC
)
) ON [PRIMARY] TEXTIMAGE_ON [PRIMARY]
//Insert first 'root', which is technicall a child without a parent
INSERT INTO [Entity]
([Name]
,[Description]
,[Hierarchy])
VALUES
('Root A'
,hierarchyid::GetRoot().GetDescendant(NULL,NULL))
//Create the second 'root'
INSERT INTO [Entity]
([Name]
,[Hierarchy])
VALUES
('Root B'
,hierarchyid::GetRoot().GetDescendant((select MAX(hierarchy) from entity where hierarchy.GetAncestor(1) = hierarchyid::GetRoot()),NULL))
Now if you select all rows from the table, you see:
SELECT [ID]
,[Name]
,[Hierarchy],
[Hierarchy].ToString()
FROM [Entity]
ID Name Hierarchy (No column name)
1 Root A 0x58 /1/
2 Root B 0x68 /2/
I'm not sure if this would be recommended practice but conceptually it allows you to have multiple roots, as long as you consider the 2nd level in the tree as the root