Removing Duplicate Data Based On Conditions

Currently trying to solve a niggling issue with removing duplicate row data from competition entry databases using the following code...

Code:
Sub remove_duplicates()

Dim lastrow As Long
Dim i As Long, j As Long

On Error GoTo errhandler

Columns("A:G").Sort Key1:=Range("A2"), Order1:=xlAscending, Key2:=Range("B2") _

[Code] ......

errhandler:
'MsgBox Err.Description
End Sub

What I need to happen is for duplicates to be based on both the name and email matching (columns A + B) and then remove all duplicate rows, whilst ignoring but preserving the content in columns C,D,E and F on those rows that remain. Column G contains either "Yes" or "No" depending on a users membership status. Sometimes duplicated data will contain identical name and email address but have different membership status in column G where someone was logged in when submitting a form the first time, and then logged out the next (potentially a mixture of the two over multiple times). Therefore I need any duplicate users who have different values in the G column throughout the database to be detected by the macro and we must ensure that the single row retained in the final non-duplicated list displays "Yes" in column G for that user.

Duplicate users who are always "yes" or always "no" work fine, its only the ones with mixed values for column G that need to be analysed and their return forced as Yes in the final list.

The final list should preserve the original column structure and thus be only one row per unique user.

The above code appeared to work the first time with a sample file, but when running the macro on one of my full size lists (3k - 5k rows) it returns 2 rows for each user who started off as a mixture of Yes/No... ?

Both rows will say yes in column G (so that part is correct!) but I have to re-run the macro again on the same dataset for it to remove the duplicate yes rows to get the list as it finally should be!

So I can achieve the result by running it twice, but this is a bit of a cheating way to do it and also the next stage after this involves something similar across multiple worksheets so I need something rock solid to build upon!