DBI::ProfileDumper is a subclass of DBI::Profile which dumps profile data to disk instead of printing a summary to your screen. You can then use dbiprof to analyze the data in a number of interesting ways, or you can roll your own analysis using DBI::ProfileData.

USAGE

One way to use this module is just to enable it in your $dbh:

  $dbh->{Profile} = "1/DBI::ProfileDumper";

This will write out profile data by statement into a file called dbi.prof. If you want to modify either of these properties, you can construct the DBI::ProfileDumper object yourself:

  use DBI::Profile;
  $dbh->{Profile} = DBI::ProfileDumper->new(
                        Path => [ '!Statement' ]
                        File => 'dbi.prof' );

The Path option takes the same values as in the DBI::Profile manpage. The File option gives the name of the file where results will be collected. If it already exists it will be overwritten.

You can also activate this module by setting the DBI_PROFILE environment variable:

  $ENV{DBI_PROFILE} = "!Statement/DBI::ProfileDumper";

This will cause all DBI handles to share the same profiling object.

METHODS

The following methods are available to be called using the profile object. You can get access to the profile object from the Profile key in any DBI handle:

  my $profile = $dbh->{Profile};

$profile->flush_to_disk()

Flushes all collected profile data to disk and empties the Data hash. This method may be called multiple times during a program run.

$profile->empty()

Clears the Data hash without writing to disk.

DATA FORMAT

The data format written by DBI::ProfileDumper starts with a header containing the version number of the module used to generate it. Then a block of variable declarations describes the profile. After two newlines, the profile data forms the body of the file. For example:

  DBI::ProfileDumper 1.0
  Path = [ '!Statement', '!MethodName' ]
  Program = t/42profile_data.t

  + 1 SELECT name FROM users WHERE id = ?
  + 2 prepare
  = 1 0.0312958955764771 0.000490069389343262 0.000176072120666504 0.00140702724456787 1023115819.83019 1023115819.86576
  + 2 execute
  1 0.0312958955764771 0.000490069389343262 0.000176072120666504 0.00140702724456787 1023115819.83019 1023115819.86576
  + 2 fetchrow_hashref
  = 1 0.0312958955764771 0.000490069389343262 0.000176072120666504 0.00140702724456787 1023115819.83019 1023115819.86576
  + 1 UPDATE users SET name = ? WHERE id = ?
  + 2 prepare
  = 1 0.0312958955764771 0.000490069389343262 0.000176072120666504 0.00140702724456787 1023115819.83019 1023115819.86576
  + 2 execute
  = 1 0.0312958955764771 0.000490069389343262 0.000176072120666504 0.00140702724456787 1023115819.83019 1023115819.86576

The lines beginning with + signs signify keys. The number after the + sign shows the nesting level of the key. Lines beginning with = are the actual profile data, in the same order as in DBI::Profile.

Note that the same path may be present multiple times in the data file since format() may be called more than once. When read by DBI::ProfileData the data points will be merged to produce a single data set for each distinct path.

The key strings are transformed in three ways. First, all backslashes are doubled. Then all newlines and carriage-returns are transformed into \n and \r respectively. Finally, any NULL bytes (\0) are entirely removed. When DBI::ProfileData reads the file the first two transformations will be reversed, but NULL bytes will not be restored.

AUTHOR

Sam Tregar <sam@tregar.com>

COPYRIGHT AND LICENSE

Copyright (C) 2002 Sam Tregar

This program is free software; you can redistribute it and/or modify it under the same terms as Perl 5 itself.