When you select the Normalize String check box, DQS just
ignores the special characters in your source value. DQS
internally replaces the special characters either with a null or space to ensure better string comparison for data quality processing. These internal replacements of special characters do not have any impact on the processed or output data.
If you want to find out special characters in your source data, you could use the regular expresssions in a domain rule to flag values containing special characters, and then clean them up interactively using DQS. For example, you can use the following domain
rule to flag values that contain a special character or extra spaces between two words in a value:
Value matches regular expression = ^[a-zA-Z0-9]+([\s][a-zA-Z0-9]+)*$