Handling AddNew Key Violations ADODB.Recordset - excel

Question

My previous Title caused a bit of confusion there...updated
I've encountered this problem recently while using some VBA to bridge Excel and Access.
So I have a spreadsheet containing a table that I need to import to an Access database.
The table is something generic like the following.
EmployeeNumber Unused_Field2 Unused_Field3
1 ### ###
2 ### ###
3 ### ###
The unique key in Access is set to be the EmployeeNumber.
My VBA Code in Excel looks like the following:
Sub test()
Dim con As ADODB.Connection
Dim rst As ADODB.Recordset
strcon = "Provider=Microsoft.Jet.OLEDB.4.0;Data Source=c:\temp\mydb.mdb;"
strsql = "SELECT * FROM Table1"
Set con = New Connection
Set rst = New Recordset
con.Open strcon
rst.Open strsql, strcon, adOpenStatic, adLockOptimistic
For i = 0 To n
On Error GoTo Errhdl
rst.AddNew Array("Field1", "Field2", "Field3"), Array(Range("A" & i), Range("B" & i), Range("C" & i))
On Error GoTo 0
Next
Exit Sub
Errhdl:
Debug.Print "Record" & i & "caused an error"
Resume Next
End Sub
Unfortunately the data is not of high quality and I often would have duplicate values causing a key violation all the time.
While I though Resume Next will clear the Error allowing another "AddNew" to run, it did not.
All the subsequent entries after one key violation will return the same Error.
So my question is as follows:
Is there away to clear the Error associated with a recordset? ( For the record I've tried to get the "Error Collection" and used
the method Errors.Clear. That didn't do the trick. The Error Collection is a property specific to the ADO object - MSDN Errors Collection)
Can this be done without closing and re-opening the recordset?
Please let me know if more clarification is needed!

Answer 1

I would not use error handling (handling ADO errors is tricky) but use Find to check each time that the key does not already exist, before performing the insert. If the key is an indexed field (as your description suggests) then you might consider Seek which is more efficient for a large dataset. (However, Seek does not work with a client cursor -adUseClient.)
Here is a code fragment that I found that outlines the steps:
If Not .EOF Then
.MoveFirst
.Find "TPItemNbr='" & m_TPItemNbr & "'", , adSearchForward
End If
If .EOF Then
.AddNew
!TPItemVendorID = m_TPItemVendorID
!TPItemNbr = m_TPItemNbr
!TPItemEUOM = m_TPItemEUOM
!TPItemUOMFactor = m_TPItemUOMFactor
!TPItemPUOM = m_TPItemPUOM
!TPItemDescription = m_TPItemDescription
!TPItemUnitCost = m_TPItemUnitCost
!TPItemUnitLabor = m_TPItemUnitLabor
.Update
That is, perform a Find each time within the loop, if EOF (end-of-file) is true then the key isn't already in the table, so an insert can be performed.
Added in response to further information about the compound key.
I would create a Command object and use Execute. Yes, it requires creating a string, but you can trap and ignore errors for key violations.
The alternatives could be to
Run a separate SQL-statement to obtain a recordset of the list of duplicates
Loop through this storing the values in an array
Perform the AddNew each time checking the array
This seems messy to me, particular the aspect of searching the array.
Pursue the multi-find method linked in a comment.

Answer 2

So I found a solution if a Recordset.AddNew encounters and error.
The trick is to use CancelUpdate - MSDN
One can also use the Status property to check if the operation was successful.
a sample code would be:
Sub test()
Dim con As ADODB.Connection
Dim rst As ADODB.Recordset
strcon = "Provider=Microsoft.Jet.OLEDB.4.0;Data Source=c:\temp\mydb.mdb;"
strsql = "SELECT * FROM Table1"
Set con = New Connection
Set rst = New Recordset
con.Open strcon
rst.Open strsql, strcon, adOpenStatic, adLockOptimistic
For i = 0 To n
On Error GoTo Errhdl
rst.AddNew Array("Field1", "Field2", "Field3"), Array(Range("A" & i), Range("B" & i), Range("C" & i))
On Error GoTo 0
Next
Exit Sub
Errhdl:
Debug.Print "Record" & i & "caused an error"
If rst.Status <> 0 Then
rst.CancelUpdate
End If
Resume Next
End Sub