SQL Server: How to customize Merge When Matched area - tsql

Question

See my code where I insert or update a table from XML using merge statement.
drop table if exists Student
Declare #Data xml
set #Data=
'<Root>
<Student>
<Name>Rakesh</Name>
<Marks>80</Marks>
</Student>
<Student>
<Name>Mahesh</Name>
<Marks>90</Marks>
</Student>
<Student>
<Name>Gowtham</Name>
<Marks>60</Marks>
</Student>
<Student>
<Name>Manoj</Name>
<Marks></Marks>
</Student>
</Root>'
create table Student (
Name varchar(10),
Marks int
)
insert into Student values
('Rakesh',90),
('Mahesh',80),
('Jack',80),
('Manoj',57)
DECLARE #archive TABLE
(
ActionType varchar(10),
Name varchar(10),
Marks int
);
Merge into Student as Trg
Using (select d.x.value('Name[1]','varchar(20)') as Name ,
d.x.value('Marks[1]','int') as Marks from
#data.nodes('/Root/Student')as d(x)) as Src
on Trg.Name=Src.Name
When Matched Then update set
Trg.Marks=Src.Marks
when not matched then
insert (Name,Marks) values (Src.Name,Src.Marks)
OUTPUT
$action ,
inserted.*
INTO #archive;
I want when Name match then UPDATE will perform but when Name match but Marks is empty then that records will be deleted based on name matched.
so tell me how could I customize this.
see my XML there is one student named Manoj whose marks are empty then that records will be removed from the table based on name matched.
How could I mention a condition in Merge that when Name matched and marks not empty then update and when Name matched but marks are empty then that records will be removed from the table.
Can i use Multiple matched condition ? if yes then it will be possible.
please guide me on how to achieve this. thanks

Answer 1

You should be able to use multiple WHEN MATCHED parts each with a clause search condition. One search checks that Src.Marks is not zero and the other checks that it was.
I've also put a clause search condition on the NOT MATCHED part as I would guess you don't want to be inserting people that have no marks either.
MERGE INTO Student as Trg
USING (
SELECT
d.x.value('Name[1]','varchar(20)') AS Name ,
d.x.value('Marks[1]','int') AS Marks
FROM
#data.nodes('/Root/Student') AS d(x)
) AS Src
ON Trg.Name=Src.Name
WHEN MATCHED AND Src.Marks <> 0 THEN
UPDATE
SET
Trg.Marks=Src.Marks
WHEN MATCHED AND Src.Marks = 0 THEN
DELETE
WHEN NOT MATCHED AND Src.Marks <> 0 THEN
INSERT (Name,Marks)
VALUES (Src.Name,Src.Marks)
OUTPUT
$action,
inserted.*
INTO #archive;