> From: Adam Musch <ahmusch@xxxxxxxxx> ... > The only way I can think of is with a block dumps of the temporary > segments in use by the other session, assuming *that's* even > possible. It might get the data, but unscrambling the egg ... I agree. The data is there except you don't know the starting and ending offsets. You can easily test this by alter system dump tempfile [tempfile#] block min [1] block max [2] where [tempfile#] is segfile# in v$tempseg_usage minus db_files, [1] is [segblk#] and [2] is a few numbers above it. (You may narrow it down to one single block.) Alternatively, oradebug set...pid oradebug dump heapdump 1025 also gives you the data mingled with other things. The only use of this dump (from file or memory) is when you already know most of the data and miss just a little. E.g, you know the data is the string '1-800-123-4?67' except the single digit marked as ?. Yong Huang -- //www.freelists.org/webpage/oracle-l