I have a table with two string type columns named IL and CL. I have to compare both the strings character by character and wherever there is a question mark in column CL string, I have to replace it with the character found in the other column (IL), in the exact same position.
For example, the CL string after comparing with IL and replacing the question marks would be the New CL as shown below
| CL | IL | NEW CL |
|---|---|---|
| ???-123201-000000-000-??? | 104-234561-644221-123-947 | 104-123201-000000-000-947 |
I have the below code working, where there is a while loop inside of a cursor going through each character of the string, concatenating into a new string, and then doing an update at the end with the new value. However, performance with this piece of code is horribly slow as the table has 100K+ records and it's looping through each row 25 time (string length). I am looking to see if perhaps there is a way to rewrite this logic set based for performance improvement.
declare CharC insensitive cursor for select ID, IL, CL from AcctStrings where CL like '%?%'open CharCfetch next from CharC into @ID, @IL, @CLwhile @@fetch_status = 0 begin set @NewCL = '' set @i = 1 while @i <= 25 begin set @TestChar = substring(@CL,@i,1) set @OtherChar = substring(@IL,@i,1) if (@TestChar = '?') begin set @NewCL = @NewCL + @OtherChar end else set @NewCL = @NewCL + @TestChar set @i = @i + 1 end update AcctStrings set CL = @NewCL where ID = @ID end fetch next from CharC into @ID, @IL, @CLenddeallocate CharC