The described bypass is most consistent with using an in-line comment technique. Many simplistic SQL injection filters look for specific “blocked” keywords such as SELECT, UNION, OR, or DROP as contiguous strings. If the tester breaks a keyword apart by inserting tokens the database will ignore—most commonly comment delimiters—the filter may fail to match the forbidden keyword, while the SQL parser still reconstructs the intended meaning. For example, inserting comment markers inside a keyword can cause the application-layer filter to miss the pattern, yet the database can still process the statement depending on the DBMS and parsing rules. This is why comment-based obfuscation is a common SQLi evasion approach against weak signature filters.

The scenario’s wording “broken apart with unexpected symbols” maps well to comment markers such as /**/ (or other DBMS-specific comment forms). These characters look “unexpected” to a naïve filter, but to the SQL engine they are treated as comments/whitespace and effectively ignored during parsing, allowing the attacker’s intended keyword to be interpreted correctly.

Why the other options are less aligned:

String concatenation (A) is an obfuscation approach where attackers build keywords/strings using concatenation operators or functions (DBMS-specific). The scenario emphasizes splitting keywords with symbols that the DB interprets correctly as separators/ignored content, which more directly matches comment insertion.

Hex encoding (B) encodes parts of a payload (strings, sometimes function names) in hexadecimal form. That is a different technique than splitting keywords with inserted symbols.

Null byte (D) is historically used to terminate strings in certain contexts (more common in older file/path handling bugs) and is not the primary technique for bypassing keyword filters in modern SQL parsing.

Therefore, the SQL injection evasion technique is C. In-line Comment.