SplitOn column is needed twice in the query to work properly

Hello,

I'm using Dapper 2.1.66 with SQL Server and Net 8.
I noticed a weird thing using the SplitOn feature where it doesn't work as expected.

I created a basic test case to demonstrate the issue.
I have created 2 basic tables with just a few columns in each:

• TestParent: TestParentId, TestParentNm
• TestChild: TestChildId, TestParentId, PermissionId

I use this query to pull the data:

Select p.TestChildId, p.PermissionId, p.TestParentId, r.TestParentNm 
From TestChild p inner join TestParent r ON p.TestParentId = r.TestParentId

As shown in the Dapper SplitOn example page, I only select the join column once (see: https://www.learndapper.com/relationships#dapper-spliton)

This gives this result (in green, child columns, in blue, parent columns)

Now, in my code, when I want to query the data and populate my objects, I use this code:

            var query = "Select p.TestChildId, p.PermissionId, p.TestParentId, r.TestParentId, r.TestParentNm " +
                "From TestChild p inner join TestParent r ON p.TestParentId = r.TestParentId";
            var itemList = (await dbDapperHelper
                .GetDapperConnection()
                .QueryAsync<TestChild, TestParent, TestChild>(
                query, (child, parent) => {
                    child.TestParent = parent;
                    return child;
                },
                new { },
                splitOn: "TestParentId"))
                .ToList();

GetDapperConnection() is a helper method that returns the SqlConnection

Ok, now the weird thing.
When I run this code, I'm able to populate "TestParent" with the parent object, but the TestParentId of the child object remains 0

This shows TestParentId = 0

But here you can see that the TestParent object is populated, and has a TestParentId of 4.
I would expect the TestParentId in the child to be 4 too.

Now, if I modify the SQL query to output the TestParentId column twice, it works as expected

This strikes me as weird, since the example in the official Dapper documentation shows that the splitOn column needs to be in the output columns only once.

Is there something in my set up that is wrong?
Let me know if you need more details

Thanks

Start at Dapper's QueryAsync multi-mapping path and the splitOn handling, then reproduce the supplied SQL Server example with the TestChild and TestParent objects. Compare behavior when TestParentId is selected once versus twice; done when the documented single split column preserves the child's TestParentId, with a regression test covering the case.