Finding Duplicates Regardless Of Special Characters - Help Plz

I am trying to move distinct data from one table to another based on two columns regardless of special characters such as : ( ) ; ' " / - _ = >< etc. The two columns that are the defining factors that I need to match as duplicates are Col1 & Col3, the rest I want to get the maximum data from each column for that row (I hope this made sense).

Here is what I have tried, but it does not seem to work:

Code:


select
ID,
Col1,
max(Col2)as Col2,
Col3,
max(Col4)as Col4,
max(Col5)as Col5,
max(Col6)as Col6,
max(Col7)as Col7,
max(Col8)as Col8,
max(Col9)as Col9
where patindex('%[^-:]%',Col1) = 0 AND patindex('%[^-:]%',Col3) = 0
into Newtable
from dbo.CLEANING_bk
group by Col1,Col3



Basically if there are 10 rows of duplicate data based on Col1 & Col3 (regardless of special characters), I want to move the distinct info to a new table with the maximum amount of info from the other columns of the duplicate rows.

If row 3, 4,5, & 6 are all considered duplicate and Row3/Col2 has info in it that row 4 & 5/Col2 does not, I want to combine it with the single row I export to the new table retaining as much information from the rows of the duplicates as possible.

Can anyone help and let me know why my script is not wanting to play nice?