Good question – I ended up writing a blog entry about it:

http://blog.tanelpoder.com/2009/11/04/detect-chained-and-migrated-rows-in-oracle/

Neat Compilation I should say.
Wanted your help to know if a table is suffering from row migration from any block dump? [ the reason I want to know this from dumps is that I work on a Datawarehouse database of size 200TB
and tables on it are very big ..to the tune of 500GB to 1 TB, so its impossible to do a ‘analyze table list chained rows into chained_rows’ for the following reasons:
(a) analyze statement…cannot be parallelized , only dbms_stats can be parallezied, so I cannot accelerate this activity.
(b) also tables undergo batch jobs every now and then so I would like to keep this table free of any routine maitenace activity.

Thank you for your invaluable blogs as always.

Otherwise, I think there’s a metalink note about it (search for “reading 10046 trace”) and a google search returned few hits with some details like this one:

http://oradbatips.blogspot.com/2007/06/tip-48-reading-10046-trace.html

Could you please point me to a good tutorial to understand 10046 trace file in-depth.

Thanks in Advance.

I think a good name for PARSING IN CURSOR is CURSOR ATTRIBUTES or ATTRIBUTES FOR CURSOR. I had toyed with the name TEXT FOR CURSOR but that is too specific given the recursive depth and authentication attributes given.