We have an COTS software that is sending a bogus SQL statement to the
Oracle RDBMS (18.104.22.168). The error message (ORA-00933: SQL command not
properly ended) doesn't give us a clue as to what is really wrong because
it appears to be a data item rather than a syntax error. Because it appears
to be a literal, it doesn't make any difference to the SQL area.
Is there any way to "sniff" what the SQL is that comes through the TNS? I
know of tracing, but I'm not sure how or at what level might catch this.