ok, the complete workflow is a little bit complicated. we are using a
workflow engine and the newsletter generator is one step of three. the
first cleans the statistics data, the second generates the new data and
the third is the one which generates the mails.
Put this in between each line and see where your memory is jumping the most:
error_log('in file ' . __FILE__ . ' at line ' . __LINE__ . '; memory usage is ' . number_format((memory_get_usage() / 1024), 4) . ' kb');
Once you've worked out the biggest jumps, fix them, then work on the next one and so on.
It's going to be a lot quicker doing something like that over asking us to help you work out thousands of lines of code.
--
Postgresql & php tutorials
http://www.designmagick.com/