Openworld Day 2

Yesterday I went to a presentation on histograms in 12c.  A number of good improvements made there by Oracle coming in 12c.

Particular emphasis seems to be placed on some of the short comings of height balanced histograms.  In 12c, we will have TOP-N histograms (where the "top scorers" get their own frequency bucket, and the other values get collapsed together.  Similarly, there is a hybrid form of height balanced histogram.  Lots of playing around to be done there.

The limit on histograms has been bumped up to 2048, although you have to ask the question – if you have 2000 special values, then maybe, you don’t actually have *any* special values 🙂

Later in the day, I went to Bryn Llewellyn’s PLSQL enhancements talk.  PLSQL features in new releases is always an interesting one for me in that its such a mature product, so does an absence of "headline" new features mean that its no longer cared for, is that it is functionallty complete ?  You never can tell.

Better SQL and PLSQL interaction seemed to be the main thrust.  You can have functions within a WITH clause, as a means of being able to leverage the ‘procedural’ elements of PLSQL without having to recast all of that into SQL expressions. 

There is also a new pragma to make PLSQL that it called from SQL even faster.  I think this will get a lot of takeup from those who move to 12c, because calling PLSQL from SQL has become so commonplace nowadays.

UTL_CALL_STACK is a now package introduced, which (finally) gives some structure to the call stack that we all use the parse out of DBMS_UTILITY.