Cursor.executemany() fast-path parsing for INSERT ... VALUES ... ON DUPLICATE KEY UPDATE did not accept MySQL 8.0.19+ alias syntax (VALUES (...) AS alias), so these statements could miss bulk-insert optimization. This updates the parser to recognize the newer syntax while preserving existing behavior.

• Regex update in pymysql/cursors.py

  • Extend RE_INSERT_VALUES suffix parsing to accept:
    • optional AS <row_alias>
    • optional AS <row_alias>(<col_alias>, ...)
    • followed by optional ON DUPLICATE ...
  • Keep existing INSERT|REPLACE ... VALUES (...) grouping semantics used by executemany().

• Compatibility preserved

  • Existing legacy upsert style using VALUES(col) in ON DUPLICATE KEY UPDATE remains supported.

• Focused coverage in pymysql/tests/test_cursor.py

  • Add regex assertions for:
    • VALUES (...) AS new ON DUPLICATE KEY UPDATE ...
    • VALUES (...) AS new(c1,c2,...) ON DUPLICATE KEY UPDATE ...
    • legacy ON DUPLICATE KEY UPDATE ... VALUES(...)

m = pymysql.cursors.RE_INSERT_VALUES.match(
    "INSERT INTO t1 (a,b,c) VALUES (%s,%s,%s) AS new "
    "ON DUPLICATE KEY UPDATE c = new.a + new.b"
)