Re: SQL:200n Summary of major changes?

"Tibor Karaszi" <tibor_not_pressed_ham_.karaszi_at_cornerstone.se> wrote in message news:<xCTm9.345$MV.15560_at_newsc.telia.net>...

> I'm looking for some summary of changes in the proposal for SQL:200n
> (compared to SQL:1999). I know that the current draft is available from
> http://www.wiscorp.com/SQLStandards.html, but the document only explicitly
> lists deprecated features, not new elements.

First of all, rather than wiscorp.com, I'd point to the more official site of the standards committees where .pdf downloads are available: http://sqlstandards.org/SC32/WG3/Progression_Documents/Informal_working_drafts/ ... even though it's often down (see the complaints on the committees' mailing lists at sql-99.org). Incidentally, if you look at the agendas and postings on that site, you'll see the the term "SQL:200n" is being gradually replaced by "SQL:2003".

I can't call this a summary because I only noticed what I cared about, so far:
(1) More aggregate functions, e.g. VAR_xxx and STDDEV_xxx.
(2) More scalar functions, e.g. CEILING.
(3) Window partitioning.
(4) A recognition that all character sets are a subset of Unicode.
(5) Disappearance of BIT and BIT VARYING.
(6) Parts for SQLJ and XML.
(7) Part for information_ and definition_schema, shifted from
foundation.
(8) SEQUENCE + IDENTITY, though with characteristics that look odd to
me.
(9) More enduring "session characteristics" settings in Foundation and
in CLI.
(A) More features in the OLAP package.
(B) "Translations" have become "transliterations" -- a clarification.

The item that I'd pick out for special attention is is Part 14 - XML-Related Specifications (SQL/XML). With a clear description of what object/relational data looks like physically, we have a common storage and transfer format for exchanges between DBMSs, or to and from objects in other programming languages,
especially Java. I suppose that much of the change can be called housekeeping or incremental, but if everyone actually complies with SQL/XML (so far nobody does), the world will become a better place.

For the last few weeks I've been promising: my article on standards adherence among major DBMS vendors will appear online soon at http://dbazine.com. I apologize for the flexibility of the definition of the word "soon".

Peter Gulutzan
Co-Author of SQL-99 Complete, Really
Co-Author of SQL Performance Tuning
(http://ourworld.compuserve.com/homepages/ocelotsql/tuning.htm)
Received on Thu Oct 03 2002 - 17:27:17 CEST