Question

How would I go about creating a query to replace a social security number from the middle of a text block?

The table is like this:

column1     column2
11          text SSN more text

SSNs are not all formatted the same either, some are like this xxx-xx-xxxx and some are just a number string xxxxxxxx.

Answer 1

If you are wanting for example to redact SSN#’s from your table, you can make use of the PATINDEX function to find SSN#’s and the REPLACE command to convert them to something that hides the number. Here is an example (note that this code assumes there will only be one SSN# in a value at once. If there could be multiple different SSN#’s, only the first one will be found):

update
    MyTable
set
    Column2 = replace(Column2, 
        case
            when patindex('% [0-9][0-9][0-9]-[0-9][0-9]-[0-9][0-9][0-9][0-9] %', Column2) <> 0 
                then substring(Column2, patindex('% [0-9][0-9][0-9]-[0-9][0-9]-[0-9][0-9][0-9][0-9] %', Column2) + 1, 11)
            when patindex('% [0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9] %', Column2) <> 0 
                then substring(Column2, patindex('% [0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9] %', Column2) + 1, 9)
            else ''
        end,
        case
            when patindex('% [0-9][0-9][0-9]-[0-9][0-9]-[0-9][0-9][0-9][0-9] %', Column2) <> 0 
                then 'xxx-xx-xxxx'
            when patindex('% [0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9] %', Column2) <> 0 
                then 'xxxxxxxxx'
            else ''
        end)

Sample input data:

1 | text 123-45-6789 more text
2 | testing 894315466 and blah blah blah
3 | testing no ssn 348933 result

Sample Output:

1 | text xxx-xx-xxxx more text
2 | testing xxxxxxxxx and blah blah blah
3 | testing no ssn 348933 result