MS Access - Update and Inner Join with Multiple Rows - ms-access

Question

I found answer for this problem foe SQL Server but not MS Access.
Table NE:
IP | Status | Peer
-----------------------------
10.10.10.1 | 0 | null
10.20.1.5 | 0 | null
Table Peer:
IP | Peer | Status | Flap_Count
------------------------------------------------------
10.10.10.1 | 10.100.2.3 | 0 | 0
10.10.10.1 | 10.200.1.1 | 1 | 5
10.10.10.1 | 10.50.100.7 | 10 | 9
10.10.10.1 | 10.1.1.1 | 10 | 7
10.20.1.5 | 10.20.20.7 | 10 | 5
10.20.1.5 | 10.2.2.10 | 5 | 2
Selection Rule
get record from the biggest Status AND biggest Flap_Count
Desired result:
IP | Status | Peer
-----------------------------
10.10.10.1 | 10 | 10.50.100.7
10.20.1.5 | 10 | 10.20.20.7
Sometimes this query works and sometimes it fails:
Update NE
Inner join (Select * from Peer Order by Status Desc, Flap_Count Desc) Q
On Q.IP = NE.IP
Set NE.Status = Q.Status, NE.Peer = Q.Peer
Where NE.Status = 0
That query pops up a notification (for example) that 3 rows will be affected for IP 10.10.10.1. Which one will be kept?
When I change Order by Status Desc to Order by Status Asc the result does not change.

Answer 1

Need a unique identifier field in Peer table - an autonumber will serve.
Do not see that NE table is even needed to produce the desired output.
SELECT * FROM Peer WHERE ID IN (
SELECT TOP 1 ID FROM Peer AS Dup WHERE Dup.IP = Peer.IP
ORDER BY Dup.Status DESC, Dup.Flap_Count DESC, Dup.ID DESC);
But if you must UPDATE NE:
UPDATE NE
INNER JOIN(
SELECT * FROM Peer WHERE ID IN (
SELECT TOP 1 ID FROM Peer AS Dup WHERE Dup.IP = Peer.IP
ORDER BY Dup.Status DESC, Dup.Flap_Count, Dup.ID DESC)) AS Q1
ON NE.IP = Q1.IP SET NE.Status = [q1].[Status], NE.Peer = [q1].[Peer];
Unfortunately, correlated subquery can perform poorly with large dataset, even crashing.
Here is alternate approach that does not need unique ID field. It should return one record for each IP if two assumptions are true:
same record has highest values for both Status and Flap_Count for each IP
Status and Flap_Count will always be unique pairs within each IP group
SELECT Peer.* FROM Peer INNER JOIN (
SELECT Peer.IP, Max(Peer.Status) AS MS, Max(Peer.Flap_Count) AS MF
FROM Peer
GROUP BY Peer.IP) AS Q1
ON Peer.IP = Q1.IP AND Peer.Status = Q1.MS AND Peer.Flap_Count = Q1.MF;
An UPDATE action SQL cannot directly use an aggregate (GROUP BY) query as data source, however, I still don't see need for NE table.