Posted by Ted on 08/07/06 15:37

--CELKO-- wrote:
> >> Of course, as long as insist on inventing your own syntax (and that includes trying proprietary syntax from MySQL) you will have a hard time. <<
>
> Actually, it is the other way around; SELECT.. INTO is called a
> singleton select in Standard SQL and SQL Server is the invented syntax.

This illustrates one aspect of what I am finding so frustratng. I have
found very little information on exactly what the SQL standard actually
says. Most of my programming experience has involved the use of
FORTRAN, C/C++ or Java. In C++, for example, one could not go wrong
using classic sources such as Stroustrup's description of the language,
Josuttis' treatment of the STL and Lippman's treatment of the object
model in C++; and of course there are a good number of other
authorities whose books on aspects of C++ programming are of
considerable value. In my experience, though, there was not a lot of
difference among available compilers WRT standard compliance, at least
WRT core C++. The diiferences I found painful generally involved
periferal features that were little used for the longest time: a
consequence of some of these being so difficult. On any of these
issues, language lawyers would quarrel incessantly, while pragmatic
developers would instead say that their development tools are the final
arbiters of what is correct. This, in my current context, would amount
to saying that the RDBMS I am using at the time is the final arbiter of
what SQL syntax is correct. In Java, it is easier in that Sun's JRE is
the only arbiter of what is correct. ;-)

WRT "invented syntax", possibly better named as "extensions to the
standard", I have sometimes used such extensions provided by certain
tool vendors, and this was because they were very useful. In each
case, though, I knew they were extensions, and expected to have to
redevelop some parts of my code should I need to replace one suite of
development tools by another. In each of these cases, the vendor
documented these features as extensions beyond what the standard
specified.

My main problem right now is that none of my database tools relate what
they support to the SQL standard. They describe, to varying degrees of
success, what they do, or what SQL syntax they'll accept as correct,
but they generally don't talk about what is standard SQL and what is
their own extension. Therefore, my question to both of you is this:
"Has anyone, in recent years, examined what each of the common
proprietary RDBMS and each of the common open source RDBMS products
have done WRT implementing the standard, identifying where they have it
right or wrong and where they've provided extensions?" Or is it the
case that so many database developers are so locked into a specific
product they can afford to regard whatever theat product does is
defined to be correct regardless of what the standard may say, so
there'd be no interest in the sort of resource I am after?

Cheers,

Ted

[Back to original message]