Merge statement executing unmatch even though it matches - tsql

Question

I have this table
CREATE TABLE [dbo].[Book]
(
[PrimaryAuthorId] INT,
[AssociatedAuthorId] INT,
[Description] nvarchar(50),
[FromDate] DATETIME2 GENERATED ALWAYS AS ROW START HIDDEN NOT NULL CONSTRAINT [DF_Book_FromDate] DEFAULT SYSUTCDATETIME(),
[ToDate] DATETIME2 GENERATED ALWAYS AS ROW END HIDDEN NOT NULL CONSTRAINT [DF_Book_ToDate] DEFAULT CONVERT(DATETIME2, '9999-12-31 23:59:59.9999999'),
PERIOD FOR SYSTEM_TIME ([FromDate], [ToDate]),
CONSTRAINT [PK_AuthorsBook] PRIMARY KEY ([PrimaryAuthorId], [AssociatedAuthorId]),
CONSTRAINT [FK_PrimaryAuthorId_AuthorId] FOREIGN KEY ([PrimaryAuthorId]) REFERENCES [dbo].[Authors]([AuthorId]),
CONSTRAINT [FK_AssociateAuthorId_AuthorId] FOREIGN KEY ([AssociatedAuthorId]) REFERENCES [dbo].[Authors]([AuthorId]),
) WITH (SYSTEM_VERSIONING = ON (HISTORY_TABLE = dbo.[BookHistory]));
I created an spwith a merge statement which updates when combination of <PrimaryAuthorId, AssociatedAuthorId> already exist while it insert when combination doesnt exist. It does this for
2 of the combinations <PrimaryAuthorId, AssociatedAuthorId> and <AssociatedAuthorId, PrimaryAuthorId>
It has this code
declare #PrimaryAuthorId INT =16
declare #AssociatedAuthorId INT =45
declare #Description nvarchar(50)
DECLARE #Book TABLE
(
[PrimaryAuthorId] INT,
[AssociatedAuthorId] INT,
[Description] nvarchar(50),
)
INSERT INTO #Book([PrimaryAuthorId], [AssociatedAuthorId], [Description]) VALUES(#PrimaryAuthorId, #AssociatedAuthorId)
MERGE [dbo].Book D
USING (
SELECT [PrimaryAuthorId]
,[AssociatedAuthorId]
,[EndDate]
,#UpdatedBy
FROM
#AuthorAssociation
) S
(
[PrimaryAuthorId]
,[AssociatedAuthorId]
,[EndDate]
,[UpdatedBy]
)
ON D.[PrimaryAuthorId] = S.[PrimaryAuthorId] AND D.[AssociatedAuthorId] = S.[AssociatedAuthorId]
WHEN MATCHED THEN
UPDATE SET
D.[PrimaryAuthorId] = S.[PrimaryAuthorId],
D.[AssociatedAuthorId] = S.[AssociatedAuthorId],
D.[EndDate] = S.[EndDate],
D.[UpdatedBy] = S.[UpdatedBy]
WHEN NOT MATCHED THEN
INSERT
(
[PrimaryAuthorId]
,[AssociatedAuthorId]
,[StartDate]
,[EndDate]
)
VALUES
(
#PrimaryAuthorId
,#AssociatedAuthorId
,#StartDate
,#EndDate
);
MERGE [dbo].Book D
USING (
SELECT [PrimaryAuthorId]
,[AssociatedAuthorId]
,[EndDate]
,#UpdatedBy
FROM
#AuthorAssociation
) S
(
[PrimaryAuthorId]
,[AssociatedAuthorId]
,[EndDate]
,[UpdatedBy]
)
ON D.[PrimaryAuthorId] = S.[AssociatedAuthorId] AND D.[AssociatedAuthorId] = S.[PrimaryAuthorId]
WHEN MATCHED THEN
UPDATE SET
D.[PrimaryAuthorId] = S.[PrimaryAuthorId],
D.[AssociatedAuthorId] = S.[AssociatedAuthorId],
D.[EndDate] = S.[EndDate],
D.[UpdatedBy] = S.[UpdatedBy]
WHEN NOT MATCHED THEN
INSERT
(
[PrimaryAuthorId]
,[AssociatedAuthorId]
,[StartDate]
,[EndDate]
)
VALUES
(
#AssociatedAuthorId
,#PrimaryAuthorId
,#StartDate
,#EndDate
);
It works the first time it executes which inserts 2 records of the combination of the authorids
The problem is when I run the same script again, I expect it to trigger the update as there is a matching record, but it seems to give me this error on the second merge statement
Violation of PRIMARY KEY constraint PK_AuthorsBook
Why does it trigger the insert when there is a match?