Developed for use with Microsoft Purview DLP, which uses the PCRE-compatible "Boost.Regex" engine for pattern matching.

Matches US Social Security number patterns:

• Exclude all-zero area, group, or serial segment sequences: 00-XX-XXXX, XXX-00-XXXX, or XXX-XX-0000
• Exclude group numbers 666 and 9## 666-XX-XXXX or 900-XX-XXXX
• Exclude ascending and descending number sequences: 123-45-6789, 876-54-3210, etc.
• Excludes known retired SSNs: 078-05-1120 and 219-09-9999
• Allows dash, em-dash, en-dash, space, slash, tab, and "null" separators
• Boundary checks to prevent matching on telephone, credit card, and other non-SSN types.
• Excludes sequences ending with common file extensions: .pdf, .doc(x), .xls(x), .ppt(x), .zip, .jp(e)g, .png, and .log

Derived from: Comprehensive US SSN (Social Security Number) Also uses patterns from the "U.S. Social Security Number (SSN) (Nucleuz Inc)" Sensitive Information Type contained in the Microsoft Purview DLP tool.