I am new to DBMS_PROFILER. All the examples I have seen use a simple top-level procedure to demonstrate the use of the profiler, and from there get all the line numbers etc. I deploy all code in packages, and I am having great difficulty getting my profile session to populate the plsql_profiler_units with useful data. Most of my runs look like this:

RUNID RUN_COMMENT UNIT_OWNER  UNIT_NAME         SECS PERCEN
----- ----------- ----------- -------------- ------- ------
5     Test        <anonymous> <anonymous>        .00    2.1
      Profiler

5     Test        <anonymous> <anonymous>        .00    2.1
      Profiler

5     Test        <anonymous> <anonymous>        .00    2.1
      Profiler

I have just embedded the calls to the dbms_profiler.start_profiler, flush_data and stop_profiler as per all the examples. The main difference is that my code is in a package, and calls in to other package. Do you have to profile every single stored procedure in your call stack? If so that makes this tool a little useless!

I have checked http://www.dba-oracle.com/t_plsql_dbms_profiler.htm for hints, among other similar sites.